Respiratory Support Of Newborn Infants In The Camille Omar ...
179
Respiratory Support Of Newborn Infants In The Delivery Room and Neonatal Intensive Care Camille Omar Farouk KAMLIN MB BS FRCPCH FRACP A thesis submitted in total fulfilment of the requirements for the degree of Doctor of Medical Science Department of Obstetrics & Gynaecology Faculty of Medicine, Dentistry and Health Sciences University of Melbourne February 2010
Transcript of Respiratory Support Of Newborn Infants In The Camille Omar ...
Respiratory Support Of Newborn Infants In The Delivery Room and Neonatal Intensive Care
Camille Omar Farouk KAMLIN
MB BS FRCPCH FRACP
A thesis submitted in total fulfilment of the requirements for the degree of
Doctor of Medical Science
Department of Obstetrics & Gynaecology
Faculty of Medicine, Dentistry and Health Sciences
University of Melbourne
February 2010
(iii)
ABSTRACT
Background
Many newborn infants have respiratory difficulties from birth. The newborn infant’s
lung is vulnerable to injury from interventions provided in the delivery room (DR) and
the neonatal intensive care unit (NICU). Some of the possible techniques to assess,
stabilise and manage infants with respiratory difficulty whilst minimising harm have
been explored in this thesis.
Aims
The aims of this thesis were to
1. investigate the clinical utility of pulse oximetry in the DR
2. document observations on pressures and tidal volumes from a respiratory function
monitor in spontaneously breathing infants and those receiving assisted breaths
3. document difficulties encountered whilst intubating infants
4. investigate the effects of inspiratory times on the morbidity of ventilated infants in
the NICU
5. Determine the utility of a spontaneous trial of breathing to predict successful
extubation.
Methods and Subjects
1. Human studies of newborn infants born at or admitted to The Royal Women’s
Hospital, Melbourne, including observational studies in the DR and NICU, and
randomised controlled studies in the DR.
2. Bench top studies of simulated neonatal resuscitation using manual ventilation
devices and face masks
3. Meta-analysis and systematic review of the literature using guidelines as set by the
Neonatal Group of the Cochrane Collaboration.
(iv)
Results
Major results from my studies of pulse oximetry include the following: 1) The
technique of sensor application can affect the time to obtain data and the quality of
data on oxygen saturations in the NICU and DR. 2) Newborn infants not receiving any
respiratory support in the first five minutes after birth commonly have oxygen
saturations much lower than those targeted in the NICU and that the values increase
gradually with time. 3) Clinical evaluation of heart rate is unreliable and inaccurate,
whilst pulse oximetry is accurate and precise in measuring infant heart rate compared
with electrocardiography.
Major results from my studies of video recordings of neonatal resuscitation include the
following: 1) Intubation is difficult and often unsuccessful. 2) It is feasible to measure
tidal volumes in spontaneously breathing infants and those receiving assisted inflations.
3) Tidal volumes can be measured in infants with congenital diaphragmatic hernia
receiving assisted ventilation in the DR.
Major results from my bench top studies of positive pressure ventilation include the
following: 1) Using a round silicone mask, there are large leaks from masks, inflating
pressures are a poor proxy for delivered tidal volumes but improved techniques of face
mask application can be taught and retained by adult learners.
Major results from my studies in the NICU include the following: 1) Short (< 0.5
seconds) inspiratory times reduce the risk of air leak in infants with hyaline membrane
disease. 2) A three minute trial of spontaneous breathing prior during endotracheal
continuous positive airway pressure has high positive and negative predictive values for
predicting extubation success.
(v)
Conclusion
Major conclusions arising from my work include the following: 1) Pulse oximetry in
the DR can be used as an adjunct to clinical assessment to help determine the need for
providing respiratory support with objectivity. 2) Lung injury may be avoided by
monitoring tidal volumes in the DR, using short inspiratory times and improving rates
of successful extubation.
(vi)
(vii)
DECLARATION
This is to certify that
i. this thesis comprises only my original work towards the Doctor of Medical Science
except where indicated in the Preface
ii. due acknowledgement has been made in the text to all other material used
iii. the thesis is less than 100,000 words in length, exclusive of tables, maps,
bibliographies and appendices.
Camille Omar Farouk Kamlin
(viii)
(ix)
PREFACE
This thesis is a collection of published papers resulting from my efforts within the respiratory
research group at the Royal Women’s Hospital. My contribution and the contributions of my
collaborators are detailed in the individual descriptions of each study.
The work is the result of work carried out in the delivery rooms and neonatal intensive care
unit of the Royal Women’s Hospital between 2003 and 2009.
During this time period, my research activities were conducted with my colleagues Dr Colm
O’Donnell, Dr Arjan tePas, Ms Jennifer Dawson, Dr Georg Schmoelzer, Dr Liam
O’Connell, Dr Fiona Wood, Dr Kevin Wheeler, Dr Louise Owen, Professor Peter Davis and
Professor Colin Morley.
The collaborative work and resulting articles included in this thesis have also been included in
higher degrees in which my colleagues were enrolled. The tables below and overleaf identify
the studies and resultant publications that I have been involved with and data from which
were included, at least in part, in the theses submitted for examination by my colleagues.
Colleague Institution Higher Degree Year Article No
Dr Colm O’Donnell University of
Melbourne
PhD 2006 1, 2, 3, 5, 9,
15
Dr Arjan tePas University of
Leiden
PhD 2009 13, 14
Ms Jennifer Dawson University of
Melbourne
PhD 2010 4, 8
This registration date for the DMedSci was July 2009. The dates of publication for each
manuscript are detailed in the main text and in the individual papers collated in the
appendix.
(x)
Relative contribution by Dr COF Kamlin to each of the published articles included in this thesis
Position of Dr Kamlin as an author relative to total authorship
(A) Contributions To Study (B) Contribution To Writing The Manuscript
(C) Responsibility For Publication
Estimated (%) contribution by Dr Kamlin to published article
Conception & Study Design
Acquisition of Data
Analysis & Interpretation of Data
Writing Original Draft
Intellectual Input Into Revising Draft
Final Approval To Published Version
Paper 1 2nd / 4 40 Paper 2 2nd / 4 30 Paper 3 1st / 4 70 Paper 4 4th / 5 20 Paper 5 1st / 5 55 Paper 6 2nd / 6 35 Paper 7 1st / 7 60 Paper 8 2nd / 8 20 Paper 9 2nd / 4 20 Paper 10 4th / 7 20 Paper 11 4th / 7 20 Paper 12 2nd / 8 50 Paper 13 3rd / 6 15 Paper 14 3rd / 6 15 Paper 15 2nd / 4 30 Paper 16 1st / 4 75 Paper 17 1st / 2 77 Paper 18 1st / 3 75 Paper 19 1st / 5 75
(xi)
ACKNOWLEDGEMENTS
This thesis could not have been completed without the infectious energy and
enthusiasm of the respiratory research team at the Royal Women’s Hospital. Initially led
by Professor Colin Morley and latterly by Professor Peter Davis, I was helped along the
way by the vision and wit of Dr Colm O’Donnell, the gentle questioning presence of
Dr Arjan tePas and the organisational skills of Ms Jennifer Dawson. It was for the
encouragement of the whole team and the above mentioned individuals collectively,
that the work was finally completed.
Without the help and support of my family this thesis would not have come to fruition.
My parents especially my father, Mo who supported my relocation to Melbourne to
pursue my dream of becoming a researcher;; I am deeply indebted to their lifelong
support - they have been my greatest teachers. To my wife Maryam and our recent
arrival, Humza – they have been a constant source of patience, love support and
encouragement – thank you.
I would like to thank my supervisors Professors Peter Davis, Lex Doyle and Colin
Morley for their mentorship and sharing their considerable combined clinical and
research experiences with me;; they are all true leaders and exceptional teachers.
Many thanks to the babies and their parents – I have and continue to appreciate the
altruism shown by parents who provide consent for research on their newly born
children during one of the most intense times of their lives. Without them this work
would not have been possible.
Many thanks to the staff at the Royal Women’s Hospital, Melbourne for their support.
The studies presented in this thesis have been supported by a Program Grant from the
National Health and Medical Research Council and a Postgraduate Scholarship from
the Royal Women’s Hospital.
(xii)
(xiii)
ABBREVIATIONS
BPD bronchopulmonary dysplasia
bpm beats per minute
CDH congenital diaphragmatic hernia
CI confidence interval
CNRG Cochrane Neonatal Review Group
CO2
CPAP continuous positive airway pressure
carbon dioxide
DR delivery room
ECG electrocardiogram
ELBW Extremely Low Birth Weight
FiO2 fraction of
FRC functional residual capacity
inspired oxygen
H2
Hb haemoglobin
O water
HR heart rate
ILCOR International Liaison Committee on Resuscitation
IPPV intermittent positive pressure ventilation
IQR interquartile range
IT inspiratory time
min minutes
(xiv)
MV mechanical ventilation
NICU neonatal intensive care unit
NRP neonatal resuscitation program
O2
PEEP positive end expiratory pressure
oxygen
PPV positive pressure ventilation
PSANZ Perinatal Society of Australia and New Zealand
RCT randomised controlled trial
RR relative risk
RWH The Royal Women’s Hospital
SBT spontaneous breathing trial
sec seconds
SpO2
STINT
peripheral oxygen saturation
stylet for int
VLBW Very Low Birth Weight
ubation of newborn infants
VT
V
tidal volume
Te expiratory tidal volum
(xv)
TABLE OF CONTENTS
Abstract ....................................................................................................................... iii
Declaration ................................................................................................................ vii
Preface ........................................................................................................................ ix
Acknowledgements .................................................................................................... xi
Abbreviations .............................................................................................................. 13
Chapter 1. Introduction ................................................................................................ 1
1.1 An Introduction to the Work and Papers Presented in this Thesis .......................... 1
Chapter 2. Delivery Room Studies .............................................................................. 7
2.1 Neonatal Resuscitation ..................................................................................................... 72.2 Studies Of Pulse oximetry ............................................................................................... 9
2.2.1 Oxygen Saturations In The Delivery Room .............................................. 132.2.2 Heart Rate Assessment In The Delivery Room ........................................ 172.2.3 The Future – Titrating Oxygen In The Delivery Room Using Pulse
Oximetry ......................................................................................................... 222.3 Positive Pressure Ventilation In The Delivery Room ............................................... 24
2.3.1 Bench Top Manikin And Face Mask Studies ............................................ 242.3.2 Respiratory Function Monitoring In The Delivery Room ...................... 27
2.4 Spontaneous Breathing Patterns Of Newborn Infants ............................................. 292.5 Endotracheal Intubation ................................................................................................ 32
Chapter 3. Studies In The Neonatal Intensive Care Unit ......................................... 35
3.1 Time Cycled Pressure Limited Ventilation ................................................................. 353.2 Weaning Infants From Mechanical Ventilation In The NICU ................................ 37
Chapter 4. Conclusions And Future Directions ......................................................... 41
References ................................................................................................................. 45
Appendix ................................................................................................................... 53
(xvi)
Paper 1 Obtaining Pulse Oximetry Data in Neonates: A Randomized Crossover Study of Sensor Application Techniques .................................................. A1
Paper 2 Feasibility of and Delay in Obtaining Pulse Oximetry During Neonatal Resuscitation .................................................................................................. A3
Paper 3 Oxygen Saturation in Healthy Infants Immediately After Birth ............ A5
Paper 4 Pulse Oximetry for Monitoring Infants in the Delivery Room: A Review A10
Paper 5 Accuracy of Clinical Assessment of Infant Heart Rate in the Delivery Room ............................................................................................................A17
Paper 6 Accuracy of Pulse Oximetry in Assessing Heart Rate of Infants in the Neonatal Intensive Care Unit ...................................................................A20
Paper 7 Accuracy of Pulse Oximetry Measurement of Heart Rate of Newborn Infants in the Delivery Room ...................................................................A23
Paper 8 Oxygen Saturation and Heart Rate During Delivery Room Resuscitation of Infants <30 Weeks’ Gestation With Air or 100% Oxygen .............A28
Paper 9 Neonatal Resuscitation 1: A Model to Measure Inspired and Expired Tidal Volumes and Assess Leakage at the Face Mask ...........................A33
Paper 10 Assessing the Effectiveness of Two Round Neonatal Resuscitation Masks: Study ...............................................................................................A37
Paper 11 Improved Techniques Reduce Face Mask Leak During Simulated Neonatal Resuscitation: Study 2 ...............................................................A40
(xvii)
Paper 12 Ventilation and Spontaneous Breathing at Birth of Infants with Congenital Diaphragmatic Hernia ............................................................A45
Paper 13 Spontaneous Breathing Patterns of Very Preterm Infants Treated With Continuous Positive Airway Pressure at Birth .......................................A50
Paper 14 Breathing Patterns in Preterm and Term Infants Immediately After Birth ..............................................................................................................A55
Paper 15 Endotracheal Intubation Attempts During Neonatal Resuscitation: Success Rates, Duration, and Adverse Effects .......................................A60
Paper 16 Colorimetric End-Tidal Carbon Dioxide Detectors In The Delivery Room: Strenghths And Limitations. A Case Report .............................A66
Paper 17 Long versus Short Inspiratory Times in Neonates Receiving Mechanical Ventilation (Review) ...................................................................................A68
Paper 18 Predicting Successful Extubation of Very Low Birthweight Infants ......... .................................................................................................................. A101
Paper 19 A Trial of Spontaneous Breathing to Determine the Readiness for Extubation in Very Low Birth Weight Infants: A Prospective Evaluation ................................................................................................. A105
(xviii)
1
CHAPTER 1. INTRODUCTION
1.1 AN INTRODUCTION TO THE WORK AND PAPERS PRESENTED IN THIS THESIS
There have been significant improvements in the provision of care provided to newborn
term and preterm infants in the last five decades leading to dramatic improvements in the
survival rates of very immature infants. These gains have been attributed to improved
perinatal care, including the use of corticosteroids antenatally, better environments for
caring for high risk pregnancies and newborn infants in regionalised perinatal centres, use
of postnatal surfactant in respiratory distress syndrome, and the development of invasive
and non-invasive forms of ventilator support.
However, the improvements in survival appear to have plateaued recently with an
epidemiological survey showing minimal change (from 84% to 85%) in survival rates of
very low birth weight infants between two epochs of data (1995-1997 and 1997- 2002).
These were collected from the National Institute of Child Health and Human
Development Neonatal Research Network1
Rather disappointingly the rate of survival without common neonatal morbidities
(including bronchopulmonary dysplasia, intraventricular haemorrhage and necrotising
enterocolitis) between these two epochs was also static at 70%
.
1
This thesis presents 19 research papers in two themes. Firstly, investigating the respiratory
management of infants in the delivery room (DR) and secondly, investigating aspects of
ventilation in the neonatal intensive care unit (NICU).
. Improving the rates of
these morbidities in preterm infants remains a high priority for researchers in perinatal
medicine. Minimising lung injury from assisted ventilation is the major focus of this thesis.
2
An integral part of providing respiratory support is clinical evaluation. Guidelines on
neonatal resuscitation (assessment, intervention and equipment needed) are contained in
the advisory statement issued by the Neonatal Subcommittee of the International Liaison
Committee on Resuscitation (ILCOR)2
In this thesis I provide an introduction and summary for each of the research papers
published. My role and that of my collaborators are detailed in the preface and in each
chapter where the papers are discussed.
. ILCOR recommend that the newborn be assessed
clinically in the DR, with particular emphasis placed on the assessment of the infants’
respiration, heart rate and colour. The studies in the DR in this thesis are focused on the
feasibility and clinical utility of pulse oximetry as an adjunct to clinical assessment.
In keeping with the two themes, Chapter Two focuses on delivery room care of newly
born infants transitioning from an intra-uterine to extra-uterine environment. These
include studies investigating the utility of pulse oximetry in the DR and the NICU, a
literature review on using pulse oximetry in the DR. Pulse oximetry can be used to monitor
oxygen saturations (SpO2) and heart rate (HR). A major observational study of mine was
to define the normal range of oxygen saturations during the first minutes of life in infants
not receiving any assistance with breathing or supplemental oxygen. The accuracy and
precision of pulse oximetry in measuring infant heart rate is described as well as the
accuracy of clinical estimation of heart rate against the gold standard, electrocardiography.
One paper describes the changes in oxygen saturation and heart rate of infants < 30 weeks
gestation resuscitated in two historical eras. In the first cohort the standard of care was to
resuscitate infants, including those born very prematurely, with 100% oxygen. In the latter
cohort, following a change in hospital guidelines, the initial gas for resuscitation was
changed to air. The change in management was in response to accumulating evidence
demonstrating the harmful effects of oxidative injury linked to oxygen administration in
the delivery room. This study was a natural progression from the feasibility studies of pulse
oximetry and was one of the first descriptions in the literature highlighting the utility of
3
pulse oximetry in titrating the inspired concentration of oxygen during neonatal
resuscitation in the delivery room.
The provision of positive pressure inflations manually via a face mask and/or endotracheal
tube are common interventions in the DR, especially for the extremely premature infant.
The latter sections of Chapter One describe bench top studies designed to optimise
techniques of face mask ventilation and using a respiratory function monitor to document
the observations of spontaneous breathing patterns of premature infants and the tidal
volumes used to ventilate a subgroup of infants extremely vulnerable to lung injury;; infants
with congenital diaphragmatic hernia. Finally, the difficulties encountered during
intubation of infants in the DR and a case report on the limitations of exhaled carbon
dioxide (CO2
Chapter Three describes studies of assisted ventilation in the NICU. A systematic review
of long (> 0.5 secs) versus short (<0.5 secs) inspiratory times to ventilate newborn infants
is presented using the established methods of the Cochrane Neonatal Review Group
(CNRG). This is followed by two reports describing a spontaneous breathing trial (SBT) to
assist clinicians on the optimal time to extubate very low birth weight infants. If infants are
extubated too early this may lead to re-intubation, potentially causing laryngeal trauma and
leading to a longer duration of ventilation. On the other hand, if physicians are reluctant to
extubate and delay extubation, this may lead to nosocomial infections, laryngeal trauma and
may contribute to the pathogenesis of bronchopulmonary dysplasia (BPD). Using an
objective assessment of readiness for extubation such as the SBT or a combination of
pulmonary function tests may be clinically important in reducing common neonatal
morbidities. The clinical outcomes of using the SBT compared to a historical cohort are
described in the second paper.
) detectors in confirming correct placement of the endotracheal tube are
described. In addition, the results from a recently completed randomised controlled trial
(manuscript not published so not included in this thesis) comparing the use of a stylet with
no stylet to improve rates of successful endotracheal intubation are briefly described.
4
In Chapter Four, I draw conclusions from the papers presented, reflect on the nature of
the studies, the results and discuss which direction future studies may be undertaken to
minimise morbidity associated with providing assisted ventilation to the preterm infant.
(5)
The nature of the inter-relationships of the individual studies is shown in Figure 1
Figure 1. Schematic of a hypothetical patient journey of a newly born infant;; assessment of respiratory status and specific
interventions provided in the DR and NICU to assist ventilation. Superscript numbers relate to papers presented in this
thesis.
Delivery Room
Gas – oxygen / air / blenders3, 4, 8,
Patient Monitoring – pulse oximetry1, 2, 3, 4, 5, 6, 7, 8
Endotracheal intubation15 - CO2 detectors16
Interface – face mask9, 10, 11
Respiratory function monitoring of spontaneous &
assisted breaths12, 13, 14
Clinical evaluation on the need for respiratory
support2, 5, 6, 7, 8
Manual ventilation device – T Piece or self inflating
bag9, 10, 11
Mechanical ventilation
(MV): Setting the
inspiratory time17
Weaning and extubation
from MV18, 19
Intubation15
Assisted Ventilation
Birth
&
Transition
NICU
(7)
CHAPTER 2. DELIVERY ROOM STUDIES
2.1 NEONATAL RESUSCITATION
“In relation to resuscitation in the delivery room, it is now clear that it is not only
important to do the ‘right thing’, but to ask ourselves: Are we doing ‘things right’?”
Augusto Sola, Richard Deulofeut & Marta Rogido3
During the first few minutes after birth, traditional assessment of the newly born infant
is subjective and may be inaccurate. As a consequence the infant may be exposed to a
number of potentially harmful interventions or therapies including a higher
concentration of inspired oxygen than appropriate, unnecessary intubation and
ventilation, inappropriately high or low tidal volumes, and a lack of a positive end
expiratory pressure, all of which can cause lung injury.
Preterm birth is the main contributor to perinatal mortality and morbidity. The
incidence of NICU admission and duration of hospital stay are inversely related to
gestational age. There has been a great deal of research into the management of infants
born preterm once they have been admitted to the NICU. However, the number of
clinical studies investigating the management of preterm infants in the DR is limited.
Large parts of international and national neonatal resuscitation guidelines have been
based on historical dogma and consensus opinion, rather than research. It is possible
that the lungs of the most immature infants are injured during the first few minutes
after birth, well before they enter the NICU and this predisposes them to ongoing
problems.
The studies of neonatal resuscitation presented in this thesis investigated and clarified
some of the recommendations for assessment and treatment of infants in the DR. The
aim was to improve the ways in which infants are stabilised and resuscitated.
(8)
Figure 2. Research trolley with respiratory function monitor, computer, pulse
oximeter juxtaposed to a resuscitation trolley where a webcam is mounted. A flow
sensor is placed in the circuit between the manual ventilating device (here self inflating
bag) and proximal to the face mask.
In figure 2, the equipment used to conduct these studies is shown. The research trolley
housing a desktop computer installed with dedicated respiratory data acquisition and
analysis physiology software program, Spectra (Grove Medical, London, UK) is
connected to a power supply. The pressures and gas flow in ventilator circuit are
measured using the Florian (Acutronic Medical Systems, Zug, Switzerland) respiratory
function monitor. It uses a flow sensor to detect gas flow and calculates the volume of
gas passing through the sensor by integration of the flow signal. An oxygen analyser is
also connected to the ventilator circuit to measure the inspired concentration of
oxygen. The output from the Florian and analyser is transferred to the computer via an
analog-digital converter. Digital data from the pulse oximeter and the webcam are
(9)
directly transferred to the computer and integrated into the Spectra software such that
it is possible to record in real-time video images of a resuscitation with measurements
of inspired oxygen, tidal volumes, pressures delivered to the infant with continuous
monitoring of oxygen saturations and heart rate, coupled with webcam video footage.
2.2 STUDIES OF PULSE OXIMETRY
Pulse oximetry measures oxygen saturations (SpO2
These studies have investigated the use of a pulse oximeter on newly born infants in the
DR.
) and heart rate (HR) continuously
and non-invasively, without the need for calibration and correlates closely with arterial
oxygen saturation.
Pulse oximetry, a technique for monitoring HR and SpO2
Improvements in pulse oximetry technology enabled me to examine the feasibility of
using oximetry immediately after birth and how this can monitor the changes in SpO
is ubiquitous in modern
emergency and intensive care settings. Pulse oximeters were introduced into neonatal
care more than 30 years ago with the aim of avoiding periods of hypoxia and hyperoxia.
They were not routinely used in the DR to assist clinicians in the assessment of the
newly born infant at that time because of limitations of the technology.
2
Pulse oximetry is based on two principles;; firstly that oxy- and deoxyhaemoglobin differ
in their relative absorption of two wavelengths of light, red and infra-red. Using a
photo detector to measure the absorption of light emitted from a diode, the pulse
oximeter determines the amount of oxyhaemoglobin in pulsatile or arterial blood and
expresses it as a percentage of haemoglobin capable of binding oxygen, the oxygen
saturation (SpO
and HR measurements during the transition from fetal to neonatal life.
2). The second principle is that only values from pulsatile flow are used
in the final analyses and from this heart rate is determined.
(10)
In 2005, the Neonatal Subcommittee of the International Liaison Committee on
Resuscitation noted the paucity of information on SpO2 for healthy term and preterm
infants during the first minutes of life2. The optimal concentration of oxygen to use at
neonatal resuscitation was unknown and it has been suggested that pulse oximetry may
be used to titrate supplemental oxygen according to the infant’s SpO2 4 5. However
before the optimal inspired oxygen concentration could be determined, more data were
needed on the changes in SpO2
Pulse oximetry, having established itself as a tool for monitoring desaturation episodes
and low SpO
as healthy infants undergo postnatal transition.
2 values, is now being used to identify and limit the deleterious effects of
hyperoxia. Concerns about even the briefest exposure to hyperoxia have been extended
to the delivery room in recent years6 7
“Oxygen is vital, not just useful. One breath of oxygen is worth five breaths of air in
this situation.”
. Historically, the benefits of high concentrations
of oxygen during neonatal resuscitation were unchallenged, as illustrated in the
following quote:
Ainley-Walker, 19758
Although well established in the NICU, the limiting factors for using pulse oximetry in
the DR were its reliability and ability to get accurate readings in a time frame that would
be clinically useful. Historically conventional pulse oximeters took several minutes to
register any data and in the earliest clinical study evaluating a new generation
technology Masimo oximeter, a median of 2.3 minutes from birth was reported for
acquisition of data in infants less than 30 weeks’ gestation
9
My first study, designed and conducted with Dr Colm O’Donnell, was to determine the
optimal technique for placing the oximeter on an infant. We were interested in defining
the fastest method for obtaining accurate data to reflect the rapid changes in cardio-
. With assessment of the
infant and resuscitation interventions required in the first two minutes after birth, these
delays suggested pulse oximetry would be impractical.
(11)
respiratory physiology during neonatal transition. Earlier publications documenting
trends in SpO2 in the newly born infant in the DR have highlighted the difficulties
encountered in obtaining oximetry data using older conventional oximeters10
.
Paper 1
Obtaining pulse oximetry data in neonates: a randomised crossover study of sensor
application techniques. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Arch Dis
Child Fetal Neonatal Ed 2005;; 90(1):F84-F85.
Role: Patient enrolment, data collection and writing /editing the manuscript. This work
has also been included in Dr CPF O'Donnell's PhD thesis, University of Melbourne
2006
This study was designed after a discussion with Professor Neil Finer at the annual
conference of the Perinatal Society of Australia and New Zealand (PSANZ) in Sydney
(2004) who described a method of placing the sensor first on the patient before
connecting the sensor to the cable/ oximeter11. Our findings confirm that the best
method for obtaining data from the Masimo Radical pulse oximeter from stable infants
in the nursery is having the oximeter switched on and applying and securing the sensor
on the right limb before connecting the sensor to the oximeter cable11
. We focused on
reporting heart rate data as this parameter is of paramount importance for making
decisions in the DR.
(12)
Paper 2
Feasibility of and delay in obtaining pulse oximetry during neonatal resuscitation.
O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. J Pediatr 2005;; 147(5):698-699.
Role: Patient enrolment, data collection and assistance with writing manuscript. This
work has also been included in Dr CPF O'Donnell's PhD thesis, University of
Melbourne 2006
This study based on our observations from videotaping high risk resuscitations in the
DR, confirms this technique of application was effective when used on infants in the
DR12
The major limitations of using early pulse oximeters were that ambient light, blood
flow, movement artefact, decreased blood volume and venous stasis from excess
pressure on the tissues from the sensor may interfere with its ability to provide accurate
data. Following our findings from both these studies using a Masimo Radical oximeter,
this method of applying pulse oximeters in the DR is now taught to medical and
nursing personnel attending deliveries at the Royal Women’s Hospital and is being used
in many other centres around the world (personal communications).
. The reduction in the time for oximetry data acquisition was much greater than
those obtained in the NICU and further confirmed the feasibility of using this
technology in the DR. The time taken from sensor application to a digital display of
data was considerably shorter (mean difference 36 seconds) if the sensor was applied to
the infant first compared with connecting it to the oximeter then the infant. This
difference was greater than that seen in paper one where the mean difference was only
10 seconds. This study used historical controls (an earlier method of application of the
sensor) to compare with the method identified in paper one and it may be argued that
the improvements seen were a result of continuing experience gained over time using
this method. The algorithms used by the oximeters were identical and the differences
may reflect the relative difficulties of applying a sensor in a wet mobile limb of a newly
born infant compared with stable infants in the NICU, and may also reflect the changes
and fluctuations of peripheral perfusion in a transitioning infant.
(13)
2.2.1 Oxygen Saturations In The Delivery Room
Paper 3
Oxygen saturation in healthy infants immediately after birth. Kamlin CO, O'Donnell
CP, Davis PG, Morley CJ. J Pediatr 2006;; 148(5):585-589.
Role: Study design, patient enrolment, data collection and analysis, writing the
manuscript. This work has also been included in Dr CPF O'Donnell's PhD thesis,
University of Melbourne 2006
In the delivery room, the newly born infant undergoes a series of complex physiological
changes involving gradual aeration of the lungs, concurrent with circulatory changes.
The changes in oxygenation accompanying aeration of the newly born lung in healthy
newborn infants, although reported earlier, were not previously well documented and
were limited by the technology available at the time. The objective of my study was to
document trends in SpO2 immediately after birth using pulse oximetry in healthy
newborn infants in the delivery room who did not require any medical intervention.
Earlier reports described term infants only delivered vaginally13, infants who were
resuscitated or received supplemental oxygen10 14 15 or who were monitored with
technology that has been superceded10 13 16 17
Harris and colleagues placed an oximeter probe around the tendo-achilles within 15
seconds of cord clamping in 76 term infants
.
17. All the mothers received supplemental
oxygen and 44 infants were delivered by caesarean section. The mean SpO2
In 1987, House et al, published data from 100 infants weighing between 850g and
5,230g using two machines measuring functional (Nellcor) and fractional (Ohmeda)
at one and
seven minutes were 61% and 82% respectively. This was one of the first studies to
show that post ductal oxygen saturations, measured by pulse oximetry, were low and
that levels less than 82% occurred in normal infants in the first seven minutes after
birth.
(14)
saturations respectively15. Functional saturation measures haemoglobin (Hb) available
to carry oxygen whereas fractional saturation calculates a value by incorporating all Hb
types namely fetal Hb, methaemoglobinaemia and carboxyhaemoglobinaemia. In the
context of dyshaemoglobinaemia, a condition rarely seen in neonates, a normal
functional saturation reading may mislead the clinicians. Sixty two infants were
delivered by caesarean section, 42 under general anaesthesia. All infants had pre-ductal
(right upper limb) measurements and the mean SpO2
In 2002, Toth and colleagues studied 50 term infants and demonstrated from a
combination of pre and post ductal readings that healthy term infants took more than
12-14 minutes to achieve saturations >95%. They found that preductal values rose
faster than post ductal readings
at 1 and 5 minutes was 59% and
82% respectively. These data are difficult to interpret as half the infants received
supplemental oxygen. The authors said that having oximetry data was useful in
objectively judging the adequacy of resuscitative efforts and identifying cyanosed
infants. They suggested that pulse oximetry may be used as an adjunct to clinical
assessment.
13. Their findings were in general agreement with those
of Harris et al17
In the time between these early publications and my study of oxygen saturations in
healthy infants immediately after birth, there were several publications from controlled
studies and meta-analyses implicating increased mortality risk if term asphyxiated
infants were initially resuscitated using 100% oxygen compared with 21%
.
7 14 18-23. In the
Resair2 study10, Rao and Ramji published data on SpO2 values at 1, 5 and 10 minutes,
as did Vento and colleagues for term and near term asphyxiated infants resuscitated
with either air or 100% oxygen10 23 24. The striking result of these studies, especially the
latter where the resuscitators were blinded to the inspired concentration of oxygen, was
that the use of 21% or 100% oxygen for resuscitation had little effect on the change in
SpO2 in the first 10 minutes after birth. I felt it appropriate to use new generation
technology to establish normal SpO2 levels in the first few minutes of life. Having
normative data would form the basis of any recommendations on titrating the inspired
(15)
oxygen concentration with blenders in the DR, thereby potentially avoiding
unnecessary oxygen exposure and minimising harm from either hypoxia or hyperoxia.
In order to be in a position to do so, “normoxia” in the first few minutes from birth
had to be defined.
The study was conducted prospectively at the Royal Women’s Hospital and I attended
205 births of term and preterm infants and collected data from a Masimo Radical pulse
oximeter placed on the right upper limb. After exclusions for either technical reasons or
the need for resuscitation, 175 infants (121 term and 54 preterm) were recruited. I
found the median SpO2 at 1, 2, 3, 4 and 5 minutes to be 63%, 70%, 77%, 84% and
90% respectively. My findings suggested that immaturity and absence of labour slowed
the rise in SpO2 in healthy infants. Maternal analgesia and anaesthesia did not influence
the rate of rise of SpO2
Before this study was undertaken, normal values of SpO
after birth.
2 were not well described and
the appropriate target ranges for sick term and preterm infants were also unknown.
With the data I have published showing the median values and interquartile ranges in
the form of box plots one could use the 25th quartile as a cut off for providing
supplemental oxygen to those infants who fail to achieve SpO2 at this level. Of note
from these box plots are the long tails that demonstrate many healthy newborns have
oximetry values less than 70%;; the lower limit of industry calibration. Whilst the true
values cannot be verified in the newly born subject by means of blood sampling, the
trends are valuable in informing birth attendants that newly born infants’ SpO2 take
minutes to reach the “normal” adult range. Using this approach of potentially targeting
an SpO2 in a newly born infant, assumes that trends in SpO2 of healthy infants also
apply to sick preterm and term infants. More data on high risk deliveries are needed.
However, I believe the study provides data on which to base informed decisions when
attempting to avoid both hypoxia and hyperoxia, guiding clinicians on how much
supplemental oxygen to use during transition after birth. Two recent randomised
controlled studies were published assessing the feasibility of titrating oxygen
concentrations during resuscitation of preterm infants in the DR. Compared with either
(16)
100% or 90% oxygen (high oxygen), attainment of target SpO2 was more successful
using 30% oxygen compared with 21% oxygen (low oxygen) as the initial admixture25 26
.
These authors suggest that 30-40% oxygen may be the optimal concentration of gas at
the start of a resuscitation of a very preterm infant. These studies, although not
designed or powered to demonstrate safety or superiority of using a lower initial
inspired oxygen concentration compared with 100% oxygen as recommended by
national and international resuscitation guidelines, have highlighted the need for pulse
oximeters, blenders and a source of compressed air in the DR to avoid hyperoxia from
100% oxygen in the first minutes after birth.
Paper 4
Pulse oximetry for monitoring infants in the delivery room: a review. Dawson JA,
Davis PG, O'Donnell CP, Kamlin CO, Morley CJ. Arch Dis Child Fetal Neonatal Ed
2007;; 92(1):F4-F7.
Role: Literature search, analysis and contributing author. Date from this paper have
also been included in Ms Jennifer Dawson’s PhD thesis, University of Melbourne 2010
This paper is a review of the use of pulse oximetry in the DR co-authored with Ms
Jennifer Dawson with my involvement being assistance with the literature review,
analyses and a contribution to the writing of the manuscript. The paper summarises
data on the history and feasibility of pulse oximetry in the DR and compared and
contrasted the reported SpO2
values from each study. We conclude that studies need to
be designed to show that pulse oximetry in the DR actually improves patient outcomes.
(17)
2.2.2 Heart Rate Assessment In The Delivery Room
Pulse oximetry, in addition to providing data on a patient’s SpO2
, also displays pulse
rate continuously. International resuscitation guidelines recommend providing
respiratory assistance to newly born infants based on infant heart rate. Positive pressure
ventilation is recommended if the heart rate (HR) falls below 100 bpm, and external
cardiac massage if the heart rate is less than 60 bpm. Bradycardia is considered a sign of
inadequate ventilation and the maintenance of a normal heart rate is an important sign
indicating successful ventilation. Two methods of clinical assessment of heart rate are
recommended by the neonatal resuscitation guidelines;; auscultation of the heart and
palpation of the umbilical cord.
Paper 5
Accuracy of clinical assessment of infant heart rate in the delivery room. Kamlin CO,
O'Donnell CP, Everest NJ, Davis PG, Morley CJ. Resuscitation 2006;; 71(3):319-321.
Role: Study design, data collection and analysis, writing manuscript.
In this study we asked two health professionals attending the birth of infants delivered
by elective caesarean section to assess the HR (count for 6 seconds and multiply by 10)
by both auscultation and palpation of the umbilical pulse. Accuracy of both methods
was compared against the gold standard, electrocardiography (ECG). I found that
auscultation was more reliable (cord pulsations were not palpable in 20% of infants
compared with auscultation which provided a HR in all infants), and that both methods
underestimated the true HR by a range of 14 to 22 bpm. If the inaccuracies seen in our
study were to occur with infants in the delivery room, we speculate that therapies based
on clinical measurements of heart rate may be administered excessively.
Both clinical methods require the attention of one operator and can interrupt
ventilation if the heart rate cannot be heard. Heart rate measured by a pulse oximeter or
(18)
electrographic monitor provides the resuscitation team with a continuous display and
provides immediate feedback about the resuscitation. I sought to investigate the
accuracy and precision of the Masimo Radical pulse oximeter against the gold standard,
ECG.
Paper 6
Accuracy of pulse oximetry in assessing heart rate of infants in the neonatal intensive
care unit. Singh JK, Kamlin CO, Morley CJ, O'Donnell CP, Donath SM, Davis PG. J
Paediatr Child Health 2008;; 44(5):273-275.
Role: supervising medical student (JKS) project, patient recruitment, data collection
and analysis, editing manuscript
Paper 7
Accuracy of pulse oximetry measurement of heart rate of newborn infants in the
delivery room. Kamlin CO, Dawson JA, O'Donnell CP, Morley CJ, Donath SM,
Sekhon J et al. J Pediatr 2008;; 152(6):756-760.
Role: Study design, patient recruitment, data collection and analysis, drafting
manuscript
Prior to evaluating the accuracy of the Masimo Radical pulse oximeter in monitoring
HR in the DR, I first wanted to confirm the accuracy and precision of this oximeter in
stable patients in the NICU. I found that the accuracy and precision of the Masimo
were clinically acceptable and better in the NICU compared with the DR. This was not
surprising as these infants were quiet and had less variability in the heart rate.
Nevertheless, the results from this study were encouraging and highlighted a potential
(19)
role for using pulse oximetry in the DR where the infants were wet, cool, often more
active, and where rapid changes in heart rate were anticipated. A scatter plot in this
paper demonstrates the almost linear relationship between heart rate measured by pulse
oximetry with electrocardiography. Importantly, for the true bradycardias (HR <100
bpm) measured by the ECG monitor, the oximeter displayed only 26 false negative data
points from a total of 232 data points. In other words over 90% of readings were in
agreement with ECG when HR was less than 100 bpm. This study showed that the
pulse oximeter accurately identified low heart rates. I concluded that the pulse oximeter
is a very useful adjunct to clinical evaluation of newly born infants at risk of
resuscitation.
These two studies suggest that in the DR, pulse oximetry offers significant advantages
over the intermittent and inaccurate clinical assessment of HR. In addition, a pulse
oximeter probe is easier to apply than ECG leads which may damage the fragile skin of
an extremely preterm infant. Studies evaluating the accuracy and precision of the pulse
oximeter for monitoring heart rate during resuscitation are needed.
(20)
Observational Studies Of Pulse Oximetry In The Delivery Room – A Discussion
To summarise, these studies have shown that there is an optimum method of Masimo
pulse oximeter sensor application that results in the fastest display of accurate data. We
have shown that it is feasible to use pulse oximetry in the delivery room. The technique
is acceptable to the resuscitation team and provides a continuous display of both
oxygen saturations and pulse rate, allowing the effectiveness of the interventions to be
assessed. I have shown it is possible to quantify the range of SpO2 values seen in
healthy newborn infants. This could be used as a reference range to titrate the inspired
oxygen in the DR. Although the design of this study was similar to earlier reports;; my
study was one of the first to obtain preductal SpO2 data using modern oximetry
technology. However, whether these data can be used to determine the optimal SpO2
Kopotic and Lindner examined the role of the Masimo pulse oximeter in making
clinical decisions in a prospective uncontrolled observational study of extremely
preterm (< 30 weeks’ gestational age) infants born at a single centre (University
Hospital of Ulm). The resuscitation protocol at that hospital included a sustained
controlled inflation of 20-30 cm H
targets in sick infants and those born extremely premature (where normative data are
hard to gather) is yet to be determined.
2O for 15 seconds delivered by a ventilator through
a nasopharyngeal tube with the purpose of establishing sufficient pulmonary functional
residual capacity. The initial resuscitation was conducted with 100% oxygen and the
concentration was adjusted using a blender to maintain a target SpO2 within a range of
80-92%. Criteria for intubation included bradycardia (<100 bpm or an FiO2 > 0.6 at 15
minutes of age). Only 68% (15/22) of infants were studied – exclusions included major
malformations (2), second twin (3) and technical difficulties in obtaining data (2). Of
the 15 infants studied, the median gestational age and birth weight were 26 weeks and
880 g respectively. Clinical decisions were made on two infants (immediate intubation
for apnoea and no intervention because of adequate spontaneous respiration). Of the
remaining 13 infants, pulse oximetry was used as an adjunct to clinical assessment to
guide interventions;; 5 infants were commenced on CPAP because of a HR >100 bpm
and an acceptable SpO2 (>80%) and 8 were intubated for meeting the above criteria for
(21)
intubation and early rescue surfactant therapy. This was also one of the first studies to
describe weaning FiO2 in the DR with the assistance of pulse oximetry from 1.0 to a
mean of 0.4. However, the small number of infants in this study, delays in acquiring
data (median 2.3 minutes for HR and SpO2
In a later study, Finer et al, compared the use of a T piece manual ventilation device
providing positive end expiratory pressure (PEEP) with a self inflating bag (no PEEP).
Across five centres, 104 infants were randomised of which 69 had pulse oximeters to
monitor SpO
) and an apparent failure rate of acquiring
data of almost 10% were major limitations. Before a change in practice could be
recommended further evidence was needed to assess the feasibility and utility for the
routine use of pulse oximetry data in the DR.
2 and HR. Decisions to intubate and treat with surfactant in the DR were
based on both SpO2
As a result of accumulating evidence from other centres and our own series of studies,
the use of pulse oximetry in the DR has become the standard of care at the Royal
Women’s Hospital. Pulse oximeters, for neonatal use, are available in every birthing
area. Oximetry has been used to monitor patients in randomised controlled trials in the
DR providing useful data for the caregiver and researcher alike. I believe there is an
important role for pulse oximetry in the management of the transition of a high risk
infant, such as an extremely preterm infant. Pulse oximetry is essential for the safe
titration of supplemental oxygen to avoid hypoxia and hyperoxia. Compared with
clinical examination, pulse oximetry is convenient and accurate, and it has the
advantage of allowing the resuscitator to continue stabilising the infant whilst providing
continuous SpO
and HR readings.
2
In the next study, our research group report our findings on newly born premature (<
30 weeks) infants’ SpO
and heart rate data. This enables the clinician to determine the need
for and the effectiveness of resuscitative interventions without the interruptions
necessary to assess heart rate.
2 levels before and after a change in DR policy in administering
supplemental oxygen.
(22)
2.2.3 The Future – Titrating Oxygen In The Delivery Room Using Pulse
Oximetry
Paper 8
Oxygen saturation and heart rate during delivery room resuscitation of infants <30
weeks' gestation with air or 100% oxygen. Dawson JA, Kamlin CO, Wong C, te Pas
AB, O'Donnell CP, Donath SM, et al. Arch Dis Child Fetal Neonatal Ed
2009;;94(2):F87-91.
Role: Data collection, analysis and writing the manuscript. This work has also been
included in Ms Jennifer Dawson’s PhD thesis, University of Melbourne 2010
In this paper we reported our hospital’s experience after a change in clinical practice.
The change in policy of beginning resuscitation with 100% to beginning with 21%
oxygen was compared using downloaded oximetry data27. For resuscitations starting
with 21% oxygen, the new hospital guidelines stipulated that this could only be changed
to 100% oxygen if the SpO2 was <70% at 5 min of life, if the HR was <100 bpm after
60 s of ventilation or if chest compressions were initiated. Of the 106 preterm infants
started with 21% oxygen, 97 (92%) infants were given supplemental oxygen at some
point in the resuscitation. The median SpO2 for these infants at 2 and 5 min of life was
31% and 54%, respectively, and increased to 81% at 6 min of life, once 100% oxygen
was administered. Conversely, the 20 infants evaluated before the practice change
(resuscitated with 100% oxygen) had median SpO2 values of 84% and 94% at 2 and 5
min of life, respectively. The nine infants successfully resuscitated with 21% oxygen had
median SpO2 87% at 5 min of life. This study showed that whilst preterm infants
frequently require assistance with ventilation due to pulmonary immaturity, the need for
supplemental oxygen is not universal. 100% oxygen was associated with inappropriately
high SpO2 values, whereas 21% was associated with oxygen levels for the first five
minutes that sometimes did not increase above fetal levels. Although the long- term
consequences of such differences in oxygenation are not known it appears likely that
neither 21% oxygen nor 100% oxygen is optimal. This study, in addition to the
(23)
randomised trials of Wang et al25 and Escrig and colleagues26, shows that a policy of
titrating oxygen concentration levels to saturation targets in the first minutes of life is
both sensible and practical. These randomised studies have used starting points of 21%,
30%, 90% and 100% oxygen and an alternative and superior starting point to the
conventional (21% or 100%) remains to be identified. In both trials, most infants
received supplemental oxygen with an average FiO2 of 0.5 at 5 minutes. . More
normative data on the SpO2 in the immediate transition period will help define SpO2
Until we have more studies using oximetry to titrate oxygen in the DR, it remains
unclear whether the pulse oximeter offers any benefit to these infants during immediate
transition. It must be acknowledged that at the Royal Women’s Hospital, pulse
oximetry in the DR has become the standard of care without rigorous evaluation from
a randomised controlled study. It is appropriate to point out that the almost universal
use of oximetry in the intensive and special cares nurseries of the world is also not
based on randomised controlled trial data.
targets which avoid the potential side effects of both hypoxia and hyperoxia.
Whilst pulse oximetry technology has improved, with algorithms in the most modern
devices compensating for low perfusion states, ambient light and motion of the limb, it
remains to be determined whether pulse oximetry in the DR increases survival,
minimises oxygen toxicity and/or the effects of “hypoxia”, reduces or increases
unnecessary interventions or reduces the costs of care.
(24)
2.3 POSITIVE PRESSURE VENTILATION IN THE DELIVERY ROOM
2.3.1 Bench Top Manikin And Face Mask Studies
Providing adequate ventilation is a pillar to effective neonatal resuscitation. International
consensus statements and agreed guidelines from national bodies recommend giving positive
pressure ventilation (PPV) with manual devices using face masks as an interface. From recent
international surveys, there appears to be a clear preference for round cushioned silicone
masks28 29
The technique for “bag and mask” ventilation is commonly taught using a manikin. Teaching
and assessment of these techniques using manikins are a part of neonatal resuscitation
courses. In these courses, participants are taught to assess adequacy of ventilation by
assessing the degree of chest excursion. According to consensus statements, most newly born
infants can be adequately ventilated with a bag and mask although there is very little evidence
to support this
.
2. Of the few infants studied in the DR, tidal volumes conventionally thought
to be adequate for tidal ventilation were rarely delivered via a face mask30 31
. However the
tidal volumes and measurements of leak around the face mask have never been reported in
very low birth weight and preterm infants.
Paper 9
Neonatal resuscitation 1: a model to measure inspired and expired tidal volumes and assess
leakage at the face mask. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Arch Dis Child
Fetal Neonatal Ed 2005;; 90(5):F388-F391.
Role: Assistance with bench top design and work. Writing and editing manuscript. This work
has also been included in Dr CPF O'Donnell's PhD thesis, University of Melbourne 2006
In this study, Dr Colm O’Donnell and I modified a standard resuscitation manikin so that we
could measure the flow of gas passing through the face mask and the resultant tidal volume
delivered to the test lung during simulated neonatal resuscitation. In a leak free system, this
(25)
model showed that when a flow sensor (pneumotachometer) is placed between the manual
ventilation device and face mask, the volume of gas returning to mask from the manikin is a
good estimate of the tidal volume entering and leaving the lung. This model was set up with
the aim of measuring tidal volume and leak measurements noninvasively. Such data should
be useful in providing feedback to resuscitators using a manikin during training sessions with
the goal of improving PPV techniques. In addition this system would permit us to make real
time measurements of tidal volume during spontaneous breathing as infants undergo
transition independently and judge when PPV is needed to provide respiratory support for
preterm infants in the DR. Using a Florian respiratory function monitor (Acutronic Medical
Systems, Ag, Switzerland), a computer running a dedicated neonatal physiology program,
Spectra (Grove Medical, UK), and a video camera I was able to record respiratory
physiology, download oximetry data, and video-record over 170 high risk resuscitations. The
data collected from these series of observations on infant colour32, agreement between
observers on Apgar scoring33
, were published and form part of Dr Colm O’Donnell’s PhD
thesis. Physiological data obtained from recordings of spontaneously breathing infants and
those infants with antenatally diagnosed congenital diaphragmatic hernia are also presented in
the current thesis.
Paper 10
Assessing the effectiveness of two round neonatal resuscitation masks: study 1. Wood FE,
Morley CJ, Dawson JA, Kamlin CO, Owen LS, Donath S et al. Arch Dis Child Fetal
Neonatal Ed 2008;; 93(3):F235-F237.
Paper 11
Improved techniques reduce face mask leak during simulated neonatal resuscitation: study 2.
Wood FE, Morley CJ, Dawson JA, Kamlin CO, Owen LS, Donath S et al. Arch Dis Child
Fetal Neonatal Ed 2008;; 93(3):F230-F234.
Role: Study design, data analyses and editing manuscript, for both articles
(26)
In this series of bench top studies, we report our results after subjecting medical and nursing
staff to an evaluation of teaching methods and techniques of holding a face mask. In these
two studies, my involvement was helping with the design, recruitment of participants and
editing the manuscript.
In the first study, health professionals trained in neonatal resuscitation were asked to provide
positive pressure ventilation to a modified Laerdal Resuscibabe™ manikin using a T-piece
ventilator (Neopuff Infant Resuscitator™) with two masks in random order;; a standard
silicone round mask and a new Fisher and Paykel (FP) round mask. None of the participants
had used the FP mask before. The techniques used to hold the mask and the expertise of
participant were noted. We identified three common holds (stem, two point top and “OK”
hold) and found the degree of leak on average to be greater than 50% irrespective of
operator experience or technique. There was no difference in leak between the two masks.
In the next paper we identified the optimal hold for round face masks on a neonatal manikin
and investigated which teaching methods led to the best retention of a taught practical skill.
Rolling the face mask from the chin to cover the nose and mouth followed by two point top
hold in conjunction with jaw lift was identified as the optimal technique by three experienced
neonatologists. When participants were provided written instruction on technique followed
by a demonstration, face mask leak was reduced by 24%. In both studies operators were
unaware of the magnitude of the leak around the mask when providing PPV. This study has
shown that neonatal staff can improve and retain a skill when combining written instruction
with a demonstration. Such methods are already used in established neonatal life support
courses but without studies looking at face mask technique in the delivery room.
(27)
2.3.2 Respiratory Function Monitoring In The Delivery Room
Paper 12
Ventilation and Spontaneous Breathing at Birth of Infants with Congenital
Diaphragmatic Hernia. te Pas AB, Kamlin CO, Dawson JA, O'Donnell C, Sokol J,
Stewart M et al. J Pediatr 2009;;154:369-373
Role: Study design, patient recruitment, data collection and analysis, drafting and
editing of manuscript
This study investigated respiratory physiological parameters recorded during the
stabilisation of infants with antenatally diagnosed congenital diaphragmatic hernia
(CDH). Mortality and morbidity rates from this condition are related to the degree of
pulmonary hypoplasia, size of defect and the presence of associated anomalies. Recent
improvements in survival have been attributed to the adoption of gentle ventilation
strategies and encouraging spontaneous breathing in the NICU. However,
recommendations on how to manage these infants in the DR are derived solely from
expert opinions. Tidal volumes used to ventilate infants with CDH have not been
described previously.
We studied 12 infants and witnessed spontaneous breathing in 11. Although most
breathed at birth, only a small proportion of these breaths coincided with a manual
inflation. We also noted larger tidal volumes with spontaneous breaths compared with
mandatory inflations putting infants at risk of both volutrauma and insufficient aeration
of these hypoplastic lungs. The compliance of the respiratory system is affected by the
severity of pulmonary hypoplasia and fluid shifts during aeration of the lungs in the
transitional period such that having a fixed peak inflating pressure and non
synchronised inflations may put infants with CDH at risk of early ventilation-induced
lung injury. Although this study is limited by the small number of infants, the paper
highlights the feasibility of using a respiratory function monitor to titrate the PIP in
order to avoid volutrauma when ventilating high risk infants in the DR.
(28)
In this context, I am, as part of the Neonatal Resuscitation Research Team, at the Royal
Women’s Hospital, currently investigating the use of a respiratory function monitor in
the DR. In a randomised controlled study, resuscitators have the monitor either visible
or masked during PPV. If the monitor is visible, the clinician may use the information
to minimise mask leak, adjust the PIP to achieve estimated optimal tidal volumes. The
results from this RCT will be available in 2010. There is also a growing number of
experts who recognise the limitations of using a fixed peak inflating pressure as a proxy
for delivering an appropriate tidal volume during neonatal transition, with some also
encouraging manufacturers to incorporate displays of tidal volume delivered on
resuscitation trolleys34
.
(29)
2.4 SPONTANEOUS BREATHING PATTERNS OF NEWBORN INFANTS
Dr Arjan tePas and I have also used the flow sensor placed proximal to the face mask
to monitor the spontaneous breathing patterns of preterm and term infants.
Paper 13
Spontaneous breathing patterns of very preterm infants treated with continuous
positive airway pressure at birth. te Pas AB, Davis PG, Kamlin CO, Dawson J,
O'Donnell CP, Morley CJ. Pediatr Res 2008;; 64(3):281-285
Paper 14
Breathing patterns in preterm and term infants immediately after birth. te Pas AB,
Wong C, Kamlin CO, Dawson JA, Morley CJ, Davis PG. Pediatr Res 2009;; 65(3):352-
6
Role: Patient recruitment, data collection and editing manuscript, for both articles. This
work has also been included in Dr AB tePas' PhD thesis, University of Leiden 2009
These two studies explored the spontaneous breathing patterns of newly born infants in
the DR35 36. International guidelines on providing positive pressure support for infants
in the DR are based on studies conducted more than 30 years ago on a small number of
spontaneously breathing term infants. Both preterm and term infants encounter
difficulties in transitioning but the mechanisms are likely to differ and as such the
methods used to resuscitate term infants may not be applicable to preterm infants. Our
aim was to describe the respiratory patterns of very premature infants breathing
spontaneously to provide an insight on how they may aerate their lungs. This is not
previously described and may clarify for the physician which infants are likely to
(30)
respond to nasal continuous positive airway pressure (CPAP) support in the DR and
NICU.
Using physiological data recorded from the resuscitations of 12 very preterm infants
there were 792 suitable breaths for analysis. Continuous positive airway pressure
(CPAP) was provided to these infants using a Laerdal round mask attached to a
Neopuff Infant Resuscitator. The inspiratory time with a mean duration of 0.36
seconds was similar in all analysed breaths. In contrast, there were distinct patterns to
the expiratory flow waveforms;; 1) an expiratory breath hold, 2) crying, 3) grunting, 4)
panting and 5) normal flow. With the exception of normal flow and panting all the
other patterns had a prolonged expiratory phase, presumably to defend lung volumes.
Expiratory braking has been previously described as a mechanism to prevent loss of
lung volume in term37 38 and preterm infants39
In summary these studies on PPV have included bench top studies evaluating face
mask technique, focusing on measurements of leak around the face mask to assess
efficacy of ventilation. These findings will inform trainers and give encouragement to
the organisers of neonatal resuscitation programs that such skills can be taught and
retained over a period of time. The human studies have been observational. Using face
masks, I have recorded the tidal volumes generated during different patterns of
. The most frequently encountered
pattern was crying and this occurred with equal frequency in term and preterm infants
as described in the second study (paper 14). Preterm infants use significantly more
expiratory breath holds than term infants. However, the number of infants studied and
number of breaths analysed may bias the frequencies and patterns of breathing we
observed. The application of a mask on the infant’s face may influence the breathing
pattern by stimulating respiratory reflexes. Thus it would be wrong to assume that all
preterm infants have similar breathing patterns. Further research is needed to see if
recognising the prevalence and frequency of expiratory braking can guide clinicians on
how best to support the infant in the DR. This may involve increasing CPAP pressures,
synchronising manual inflations with spontaneous breaths or anticipating the need to
intubate the infant.
(31)
spontaneous breathing including the first descriptions of crying as a means of
defending lung volumes in the DR. These observations need confirmation in a larger
sample size of infants and may assist clinicians in determining the optimal method of
providing support in the first minutes after birth.
(32)
2.5 ENDOTRACHEAL INTUBATION
Endotracheal intubation of infants is a mandatory competence skill for paediatric
trainees in the UK, Australia and North America. There are fewer opportunities for
trainees to acquire this skill in the DR40 41
Paper 15
and we sought to document the number and
duration of attempts, success rates and adverse effects during endotracheal intubation
in the DR. In addition, the success rates of more experienced operators (consultant
neonatologists) have not been reported before.
Endotracheal intubation attempts during neonatal resuscitation: success rates, duration,
and adverse effects. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Pediatrics 2006;;
117(1):e16-e21.
Role: Study design, patient recruitment, data collection and analysis, editing
manuscript. This work has also been included in Dr CPF O'Donnell's PhD thesis,
University of Melbourne 2006
In this study we identified all infants where intubation was attempted and captured on a
video recording in the DR. All infants had pulse oximetry data and many of the infants
had pulmonary mechanics measured during PPV using the Florian respiratory function
monitor (Acutronic, Zug, Switzerland). A time limit on the duration of attempts is not
enforced at the Royal Women’s Hospital. Intubation was attempted a total of 60 times
in 31 infants. Success rates for residents, fellows and consultants were 24%, 78% and
86% respectively. The average duration of attempts was 38, 36 and 28 seconds for
residents, fellows and consultants respectively. A quarter of the successful intubations
took longer than 30 seconds and infant deterioration (as measured by oximetry) was
more common when HR and SpO2
With the difficulties encountered by paediatric trainees intubating infants in the DR, as
reflected by the poor success rates, I was interested in improving the methods used to
train physicians. International guidelines state the use of an introducer (or stylet) as
were low before the attempt.
(33)
being optional to assist with neonatal intubation. A stylet offers extra rigidity which
may facilitate the passage of the endotracheal tube into the laryngeal inlet. There is no
evidence to demonstrate safety or efficacy of the stylet in neonatal orotracheal
intubation.
I designed a randomised controlled trial comparing the use of a stylet with no stylet for
orotracheal int
ubation (STINT study) of infants in the DR and NICU. The study was
recently completed after 302 intubations. Operators were trainees (residents or fellows)
and intubations were randomised only if an endotracheal tube was required for
ventilatory support. Residents performed 76% of intubation attempts and the overall
success rate was 55%, much higher than we observed beforehand. This may reflect the
ongoing training provided before and during the study. We found that overall success
rates were not improved by the use of a stylet, irrespective of grade of trainee, age or
size of infant, or use of premedication. In keeping with earlier observations, the
durations of attempts were rarely within the time suggested by NRP. However, the
stylet appears to be safe. The results of the STINT study will be published in 2010.
Paper 16
Colorimetric end-tidal carbon dioxide detectors in the delivery room: strengths and
limitations. A case report. Kamlin CO, O'Donnell CP, Davis PG, Morley CJ. J Pediatr
2005;; 147(4):547-548.
Role: Direct patient involvement, literature search and writing manuscript
End tidal carbon dioxide detectors are increasingly used for verification of correct
endotracheal placement. Semi-quantitative colorimetric end tidal carbon dioxide (CO2)
detectors such as the PediCap are user friendly, correctly identifying exhaled CO2
within 5 manual inflations. In this study, we report a previously unrecognised cause of
false negative reading from such detectors. An extremely preterm infant remained
(34)
apnoeic and bradycardic after mask PPV and what was thought to be successful
intubation by an experienced neonatologist (CJM). Inflating pressures were increased
on the Neopuff Infant Resuscitator but the PediCap did not show any change in
colour. The infant was re-intubated, but in spite of the inflating pressures being
increased further, there was no change in the PediCap colour until after five very
high (unrecorded pressure) inflations from a self-inflating bag with the pop off valve
occluded. We concluded that in some very preterm infants with very stiff lungs or
airway obstruction, high inflating pressures are needed to initiate alveolar ventilation
and until then, there will be no ventilation and hence no exhaled CO2. Finer and
colleagues have also reported the utility of using the PediCap to improve face mask
ventilation with lack of colour change alerting the resuscitator to adjust mask position
or perform airway opening manoeuvres to correct for obstruction42
.
(35)
CHAPTER 3. STUDIES IN THE NEONATAL INTENSIVE CARE UNIT
3.1 TIME CYCLED PRESSURE LIMITED VENTILATION
Paper 17
Long versus short inspiratory times in neonates receiving mechanical ventilation.
Kamlin CO, Davis PG. Cochrane Database Syst Rev 2004;;(4):CD004503.
Role: Literature search, extracted and performed meta-analysis of the data, wrote the
report
Infants are commonly intubated for respiratory failure and / or surfactant therapy.
Once ventilated, physicians choose a mode of ventilation, and set an inspiratory time,
peak inflating pressure, positive end expiratory pressure (PEEP) and rate. One of my
first projects as a research fellow at the Royal Women’s Hospital was to write a
systematic review for the Cochrane Neonatal Review Group (CNRG) comparing long
inspiratory (>0.5 seconds) times with short (< 0.5 seconds) inspiratory times. The
CNRG is a voluntary group of medical, nursing and allied health professionals who
perform systematic reviews of the literature to identify risks and benefits of medical
interventions. I wrote the protocol for the review and performed this systematic review
with help from Professor Peter Davis. Data from studies identified by our search
strategy were analysed using a standardised software program (Review Manager) in
accordance with established methods of the CNRG. Although the only eligible studies
were found in the pre-antenatal corticosteroid and pre-surfactant eras and before
modern ventilation modes such as volume targeting and synchronised ventilation were
available, my review showed that a long inspiratory time was associated with an
increased risk of developing air leaks. Most of the infants studied had respiratory
distress syndrome with poorly compliant lungs and short time constants. A long
inflation time during square wave ventilation increased the mean airway pressure
applied, facilitating alveolar expansion leading to improved oxygenation. In resource-
(36)
limited settings where both antenatal corticosteroids and postnatal surfactant are not
readily available to mothers and infants, this review provides data on how clinicians
may reduce the risk of air leak. Whilst these results may not be applicable to infants
being cared for in resource rich settings with advanced ventilators, for the most
commonly used modes (assist control or synchronised intermittent mandatory
ventilation), the clinician is still required to set the inflation time. A longer inflation
time may be reasonable for infants with pathophysiology associated with long time
constants such meconium aspiration syndrome and those infants with normal lungs and
ventilated for surgery or sepsis. No studies have been performed to evaluate inflation
times during mechanical ventilation of these patients. In conclusion, given the increased
rates of air leak (Relative Risk 1.56) with 95% confidence interval (1.25, 1.94), and death
(RR 1.26 (1.00, 1.59), infants with poorly compliant lungs should be ventilated with a
short (<0.5 seconds) inflation time.
(37)
3.2 WEANING INFANTS FROM MECHANICAL VENTILATION IN THE NICU
The lung as an organ of gas exchange is reliant on a pump (rib cage, respiratory muscles
and central controller). Respiratory failure due to the inadequacy of the gas exchanger
(lung) can commonly be distinguished from failure of the pump. Whilst the
physiological criteria to intubate and initiate mechanical ventilation may vary slightly
between clinicians and centres, it is often the severity of an infant’s parenchymal lung
disease and the maturity of its ventilatory pump that determines the duration of such
support. The adverse effects of prolonged endotracheal intubation include bacterial
colonisation, sepsis, tracheal trauma and bronchopulmonary dysplasia43 44
Weaning is the process of reducing the support given by the ventilator to the infant.
Quicker weaning may reduce the risk of infants acquiring the adverse effects listed
above, but premature extubation may precipitate ventilator pump failure, especially in
the extremely low birth weight infant
45. It has been suggested that applying continuous
positive airway pressure (CPAP) following extubation preserves functional residual
capacity of the lung by reducing alveolar or lobar atelectasis and/or retained secretions
or central apnoea46. However, up to 40% of extremely low birth weight infants
(<1000g) infants fail their first extubation attempt when weaned onto nasal CPAP47
Weaning infants from mechanical ventilation can be challenging and is perceived to be
as much of an art (which comes with increasing clinical experience) than science
.
48.
Several studies have explored the utility of using predictive indices based on different
physiologic parameters of the respiratory system (pulmonary mechanics and function
tests) as an adjunct to clinical decision making49-54. However unlike adult studies, these
studies remain an investigative tool as threshold values that can differentiate between
success and failure of extubation have not evolved. In an attempt to objectify
extubation decision making I described a spontaneous breathing trial (SBT) during
endotracheal CPAP.
(38)
Paper 18
Predicting successful extubation of very low birth weight infants. Kamlin CO, Davis
PG, Morley CJ. Arch Dis Child Fetal Neonatal Ed 2006;; 91(3):F180-F183.
Role: Study design, patient enrolment, data collection and analysis, writing manuscript
In this study, once the clinicians had made a decision to extubate an infant,
physiological recordings (tidal and minute volumes, SpO2 and HR) were made during
three minutes of spontaneous breathing while still intubated. Infants who were able to
maintain their SpO2 >85% and HR >100 bpm were judged to have a successful (or
pass) SBT. The test was found to be highly accurate in predicting extubation success
and also identified infants for whom extubation was most likely to fail (a high negative
predictive value). Compared with earlier described tests/ indices, the SBT is very
simple, does not require specialised equipment, is not time consuming, and is a
pragmatic assessment of an infant’s respiratory drive. Although the SBT was associated
with high positive predictive and negative predictive values for successful extubation, it
was only used in infants who were going to be extubated. Nevertheless, as a result of
these encouraging findings, the SBT was incorporated into clinical practice at the Royal
Women’s Hospital as an adjunct to clinical decision making. Therefore, the next study I
conducted was designed to audit this change in policy on the timing and success of
extubation of very low birth weight infants.
(39)
Paper 19
A trial of spontaneous breathing to determine the readiness for extubation in very low
birth weight infants: a prospective evaluation. Kamlin CO, Davis PG, Argus B, Mills
B, Morley CJ. Arch Dis Child Fetal Neonatal Ed 2008;; 93(4):F305-F306.
Role: Study design, patient enrolment, data collection and analysis, writing manuscript
With this study, I collected data prospectively, over 13 months, from ventilated infants
in whom the SBT was performed with a view to extubating the infant based on a
successful SBT. These data were compared with a cohort of VLBW infants admitted to
our hospital in 2002, two years before the SBT became standard clinical practice. I
found that in comparison to the era when clinicians decided to extubate on clinical
grounds alone, the SBT did not lead to prolonged duration of mechanical ventilation or
a higher rate of extubation failure. There was no reduction in the rate of BPD when the
SBT was used. The study also highlighted changes in ventilator duration during the two
time periods thereby limiting the strength of recommendation for using the SBT.
Whether the use of the SBT may reduce the duration of all respiratory support, BPD
and length of stay can only be addressed in the context of a large multicenter
randomised controlled study.
(41)
CHAPTER 4. CONCLUSIONS AND FUTURE DIRECTIONS
“Since intubation and positive pressure were first recommended by Flagg in North
America in 1928 and by Blaikley and Gibberd at Guy’s Hospital in 1935, a pattern of
resuscitation has evolved based on extrapolation and assumption rather than clinical
measurement. There can be few areas of medicine where the potential benefit is so great
but which have been subjected to so little evaluation.”
A.D Milner 1991
In spite of a growing number of researchers performing studies of neonatal
resuscitation, the paucity of evidence for the equipment and techniques used allows
almost every aspect of current practice to be a potential area of investigation.
Stabilisation of preterm infants in the delivery room is a very common activity for
neonatologists around the world. Research to determine the best practice for vulnerable
infants in the first few minutes after birth should remain a high priority.
Basic scientists can assist clinical researchers by testing specific hypotheses on animal
models thereby guiding future human trials. An example of this has been the recent
completion of a randomised trial by our group in the delivery room comparing a T-
Piece Resuscitator versus a self inflating bag. This study was designed after Probyn and
colleagues in a newborn lamb model suggested that a PEEP of 8 cmH2O improved
oxygenation by maintaining end expiratory lung volumes in the creation of a functional
residual capacity55
The optimal tidal volumes, use and length of sustained inflations and standard
inflations and rate of ventilation to provide infants receiving PPV during transition to
establish FRC are not known. There is increasing acceptance that not all newly born
. The International Liaison Committee on Resuscitation does not
recommend the routine use of PEEP for newly born infants in need of PPV.
(42)
preterm infants require resuscitation but rather stabilisation during transition. A focus
of future research should examine how best to provide support to a spontaneously
breathing infant. I am confident that with close collaboration with basic scientists,
physiologists and clinicians the evidence will be generated on how best to stabilise the
most immature infants in the DR including data on long term (neurodevelopmental)
outcomes.
I am currently comparing two interfaces in a randomised controlled trial;; a single nasal
tube with a Laerdal silicone round mask. Recruitment is in progress and, at the time of
writing, 25 infants have been randomised.
Other technologies which I would like to investigate further include the use of neonatal
ventilators for providing support in the DR, use of humidified warm gases of varying
flow rates, caring for infants on a standard resuscitation trolley with a modified
incubator (Giraffe) that can be used to transport infants to the NICU. I would also like
to follow up the STINT study by investigating the utility of a video laryngoscope for
intubation of infants by paediatric trainees. I would like to use pulse oximetry in the DR
to titrate the inspired oxygen to set targets determined from centile charts our
resuscitation group have developed. Future clinical trials are required to address how
well the use of this technology in the DR improves clinical care and outcomes including
the safer and more effective use of oxygen. Such clinical studies are about to start and
can only succeed with collaborative support from other centres.
In the NICU, modifications to currently used weaning modes in combination with the
three minute SBT may increase the number of infants successfully extubated. To show
a reduction in respiratory morbidity (duration of all forms of assisted ventilation) and
length of stay, large trials with multicentre study collaboration will be necessary. For
example a randomised trial to assess the effectiveness of a spontaneous breathing trial
in reducing the extubation failure rate from 25% to 15%, with 80% power and an alpha
error of 0.05 would require a total recruitment of 500 very low birth weight infants in
both the intervention and control arms.
(43)
Summary
In summary, in this thesis I have investigated the ventilatory support neonatologists
provide to infants in the DR and in the NICU. The evidence base which we draw upon
to refine our techniques is continuing to grow. My personal priorities are to
1. Determine the best interface to use to provide manual PPV in the DR.
2. Determine the need for and duration of sustained inflation to assist in the aeration
of a lung in an infant in the DR who has either never breathed or is struggling to
breathe.
3. Use a blender and an oximeter with data on the changes in SpO2 during neonatal
transition, I would like to titrate FiO2 according to SpO2
4. Improve clinical training in neonatal resuscitation, in particular neonatal intubation.
targets judged by the best
available evidence in both preterm and term infants in the DR. These studies
should incorporate long term follow up before a protocol can be deemed safe and
effective.
5. Examine weaning protocols and further refinements of the SBT which reflect
diaphragmatic dysfunction as a cause of extubation failure.
(45)
REFERENCES
1. Fanaroff AA, Stoll BJ, Wright LL, Carlo WA, Ehrenkranz RA, Stark AR, et al.
Trends in neonatal morbidity and mortality for very low birthweight infants. Am J
Obstet Gynecol 2007;;196(2):147 e1-8.
2. The International Liaison Committee on Resuscitation (ILCOR) consensus on
science with treatment recommendations for pediatric and neonatal patients: neonatal
resuscitation. Pediatrics 2006;; 117(5):e978-88.
3. Sola A, Deulofeut R. Oxygen and oxygenation in the delivery room. J Pediatr 2006;;
148(4):564-5.
4. Kattwinkel J. Evaluating resuscitation practices on the basis of evidence: the findings
at first glance may seem illogical. J Pediatr 2003;; 142(3):221-2.
5. Saugstad OD. Resuscitation with room-air or oxygen supplementation. Clin Perinatol
1998;; 25(3):741-56, xi.
6. Saugstad OD. Optimal oxygen therapy in the newborn period. Pediatr Pulmonol Suppl
2004;; 26:112-3.
7. Davis PG, Tan A, O'Donnell CP, Schulze A. Resuscitation of newborn infants with
100% oxygen or air: a systematic review and meta-analysis. Lancet 2004;;
364(9442):1329-33.
8. Ainley-walker JC. Resuscitation of the newborn. Br Med J 1979;; 2(6204):1590.
9. Kopotic RJ, Lindner W. Assessing high-risk infants in the delivery room with pulse
oximetry. Anesth Analg 2002;; 94(1 Suppl):S31-6.
(46)
10. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the delivery room.
Indian Pediatr 2001;; 38(7):762-6.
11. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Obtaining pulse oximetry data in
neonates: a randomised crossover study of sensor application techniques. Arch Dis Child
Fetal Neonatal Ed 2005;; 90(1):F84-5.
12. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Feasibility of and delay in
obtaining pulse oximetry during neonatal resuscitation. J Pediatr 2005;; 147(5):698-9.
13. Toth B, Becker A, Seelbach-Gobel B. Oxygen saturation in healthy newborn infants
immediately after birth measured by pulse oximetry. Arch Gynecol Obstet 2002;;
266(2):105-7.
14. Saugstad OD, Rootwelt T, Aalen O. Resuscitation of asphyxiated newborn infants
with room air or oxygen: an international controlled trial: the Resair 2 study. Pediatrics
1998;; 102(1):e1.
15. House JT, Schultetus RR, Gravenstein N. Continuous neonatal evaluation in the
delivery room by pulse oximetry. J Clin Monit 1987;; 3(2):96-100.
16. Dimich I, Singh PP, Adell A, Hendler M, Sonnenklar N, Jhaveri M. Evaluation of
oxygen saturation monitoring by pulse oximetry in neonates in the delivery system. Can
J Anaesth 1991;; 38(8):985-8.
17. Harris AP, Sendak MJ, Donham RT. Changes in arterial oxygen saturation
immediately after birth in the human neonate. J Pediatr 1986;; 109(1):117-9.
18. Saugstad OD, Ramji S, Soll RF, Vento M. Resuscitation of newborn infants with
21% or 100% oxygen: an updated systematic review and meta-analysis. Neonatology
2008;; 94(3):176-82.
(47)
19. Saugstad OD, Ramji S, Vento M. Resuscitation of depressed newborn infants with
ambient air or pure oxygen: a meta-analysis. Biol Neonate 2005;; 87(1):27-34.
20. Ramji S, Rasaily R, Mishra PK, Narang A, Jayam S, Kapoor AN, et al. Resuscitation
of asphyxiated newborns with room air or 100% oxygen at birth: a multicentric clinical
trial. Indian Pediatr 2003;; 40(6):510-7.
21. Ramji S, Ahuja S, Thirupuram S, Rootwelt T, Rooth G, Saugstad OD. Resuscitation
of asphyxic newborn infants with room air or 100% oxygen. Pediatr Res 1993;; 34(6):809-
12.
22. Vento M, Asensi M, Sastre J, Garcia-Sala F, Pallardo FV, Vina J. Resuscitation with
room air instead of 100% oxygen prevents oxidative stress in moderately asphyxiated
term neonates. Pediatrics 2001;; 107(4):642-7.
23. Vento M, Asensi M, Sastre J, Lloret A, Garcia-Sala F, Vina J. Oxidative stress in
asphyxiated term infants resuscitated with 100% oxygen. J Pediatr 2003;; 142(3):240-6.
24. Saugstad OD, Ramji S, Rootwelt T, Vento M. Response to resuscitation of the
newborn: early prognostic variables. Acta Paediatr 2005;; 94(7):890-5.
25. Wang CL, Anderson C, Leone TA, Rich W, Govindaswami B, Finer NN.
Resuscitation of preterm neonates by using room air or 100% oxygen. Pediatrics 2008;;
121(6):1083-9.
26. Escrig R, Arruza L, Izquierdo I, Villar G, Saenz P, Gimeno A, et al. Achievement of
targeted saturation values in extremely low gestational age neonates resuscitated with
low or high oxygen concentrations: a prospective, randomized trial. Pediatrics 2008;;
121(5):875-81.
27. Dawson JA, Kamlin CO, Wong C, te Pas AB, O'Donnell CP, Donath SM, et al.
Oxygen saturation and heart rate during delivery room resuscitation of infants <30
(48)
weeks' gestation with air or 100% oxygen. Arch Dis Child Fetal Neonatal Ed 2009;;
94(2):F87-91.
28. O'Donnell CP, Davis PG, Morley CJ. Positive pressure ventilation at neonatal
resuscitation: review of equipment and international survey of practice. Acta Paediatr
2004;; 93(5):583-8.
29. O'Donnell CP, Davis PG, Morley CJ. Neonatal resuscitation: review of ventilation
equipment and survey of practice in Australia and New Zealand. J Paediatr Child Health
2004;; 40(4):208-12.
30. Field D, Milner AD, Hopkin IE. Efficiency of manual resuscitators at birth. Arch
Dis Child 1986;; 61(3):300-2.
31. Hoskyns EW, Milner AD, Hopkin IE. A simple method of face mask resuscitation
at birth. Arch Dis Child 1987;; 62(4):376-8.
32. O'Donnell CP, Kamlin CO, Davis PG, Carlin JB, Morley CJ. Clinical assessment of
infant colour at delivery. Arch Dis Child Fetal Neonatal Ed 2007;; 92(6):F465-7.
33. O'Donnell CP, Kamlin CO, Davis PG, Carlin JB, Morley CJ. Interobserver
variability of the 5-minute Apgar score. J Pediatr 2006;; 149(4):486-9.
34. Kattwinkel J, Stewart C, Walsh B, Gurka M, Paget-Brown A. Responding to
compliance changes in a lung model during manual ventilation: perhaps volume, rather
than pressure, should be displayed. Pediatrics 2009;; 123(3):e465-70.
35. te Pas AB, Davis PG, Kamlin CO, Dawson J, O'Donnell CP, Morley CJ.
Spontaneous breathing patterns of very preterm infants treated with continuous
positive airway pressure at birth. Pediatr Res 2008;; 64(3):281-5.
(49)
36. te Pas AB, Wong C, Kamlin CO, Dawson JA, Morley CJ, Davis PG. Breathing
patterns in preterm and term infants immediately after birth. Pediatr Res 2009;; 65(3):352-
6.
37. Kosch PC, Stark AR. Dynamic maintenance of end-expiratory lung volume in full-
term infants. J Appl Physiol 1984;; 57(4):1126-33.
38. Milner AD, Saunders RA, Hopkin IE. Is air trapping important in the maintenance
of the functional residual capacity in the hours after birth? Early Hum Dev 1978;; 2(2):97-
105.
39. Stark AR, Cohlan BA, Waggener TB, Frantz ID, 3rd, Kosch PC. Regulation of end-
expiratory lung volume during sleep in premature infants. J Appl Physiol 1987;;
62(3):1117-23.
40. Leone TA, Rich W, Finer NN. Neonatal intubation: success of pediatric trainees. J
Pediatr 2005;; 146(5):638-41.
41. Little MP, Jarvelin MR, Neasham DE, Lissauer T, Steer PJ. Factors associated with
fall in neonatal intubation rates in the United Kingdom--prospective study. BJOG 2007;;
114(2):156-64.
42. Finer NN, Rich W, Wang C, Leone T. Airway obstruction during mask ventilation
of very low birth weight infants during neonatal resuscitation. Pediatrics 2009;;
123(3):865-9.
43. Bancalari E, Sinclair J, editors. Mechanical Ventilation. New York: Oxford University
Press, 1992.
44. Avery ME, Tooley WH, Keller JB, Hurd SS, Bryan MH, Cotton RB, et al. Is
chronic lung disease in low birth weight infants preventable? A survey of eight centers.
Pediatrics 1987;; 79(1):26-30.
(50)
45. Watchko JF, Balsan MJ. Ventilatory pump failure in premature newborns. Pediatr
Pulmonol 1994;; 17(4):231-3.
46. Davis PG, Henderson-Smart DJ. Nasal continuous positive airways pressure
immediately after extubation for preventing morbidity in preterm infants. Cochrane
Database Syst Rev 2003(2):CD000143.
47. Stefanescu BM, Murphy WP, Hansell BJ, Fuloria M, Morgan TM, Aschner JL. A
randomized, controlled trial comparing two different continuous positive airway
pressure systems for the successful extubation of extremely low birth weight infants.
Pediatrics 2003;; 112(5):1031-8.
48. Sinha SK, Donn SM. Weaning from assisted ventilation: art or science? Arch Dis
Child Fetal Neonatal Ed 2000;; 83(1):F64-70.
49. Veness-Meehan KA, Richter S, Davis JM. Pulmonary function testing prior to
extubation in infants with respiratory distress syndrome. Pediatr Pulmonol 1990;; 9(1):2-6.
50. Balsan MJ, Jones JG, Watchko JF, Guthrie RD. Measurements of pulmonary
mechanics prior to the elective extubation of neonates. Pediatr Pulmonol 1990;; 9(4):238-
43.
51. Gillespie LM, White SD, Sinha SK, Donn SM. Usefulness of the minute ventilation
test in predicting successful extubation in newborn infants: a randomized controlled
trial. J Perinatol 2003;; 23(3):205-7.
52. Wilson BJ, Jr., Becker MA, Linton ME, Donn SM. Spontaneous minute ventilation
predicts readiness for extubation in mechanically ventilated preterm infants. J Perinatol
1998;; 18(6 Pt 1):436-9.
53. Khan N, Brown A, Venkataraman ST. Predictors of extubation success and failure
in mechanically ventilated infants and children. Crit Care Med 1996;; 24(9):1568-79.
(51)
54. Baumeister BL, el-Khatib M, Smith PG, Blumer JL. Evaluation of predictors of
weaning from mechanical ventilation in pediatric patients. Pediatr Pulmonol 1997;;
24(5):344-52.
55. Probyn ME, Hooper SB, Dargaville PA, McCallion N, Crossley K, Harding R, et al.
Positive end expiratory pressure during resuscitation of premature lambs rapidly
improves blood gases without adversely affecting arterial pressure. Pediatr Res 2004;;
56(2):198-204.
(53)
APPENDIX
Compilation of publications resulting from the body of work submitted for
DMED SCI
(54)
SHORT REPORT
Obtaining pulse oximetry data in neonates: a randomisedcrossover study of sensor application techniquesC P F O’Donnell, C O F Kamlin, P G Davis, C J Morley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Arch Dis Child Fetal Neonatal Ed 2005;90:F84–F85. doi: 10.1136/adc.2004.058925
Pulse oximetry may be useful during neonatal resuscitation. Arandomised crossover study was performed to determine themost efficient method of applying the sensor. Applying it tothe infant before connecting to the oximeter resulted inquickest acquisition of accurate heart rate. This techniqueshould be preferred during resuscitation.
The need for and response to neonatal resuscitation isdetermined clinically.1 Auscultation and palpation ofheart rate (HR) are subjective, intermittent, and of
questionable accuracy.2 3 Assessment of colour (a proxy foroxygen saturation, SpO2) is subjective. Pulse oximetry,routinely used in intensive care, gives continuous, accuratemeasures of HR and SpO2. Although not routinely usedduring neonatal resuscitation, it is potentially useful indetermining HR. Also, although debate continues aboutwhether air or oxygen should be used, a role for oximetry intitrating oxygen concentrations during resuscitation has beensuggested.4 Difficulties in obtaining oximetry data duringresuscitation have been reported.5 Newer generation oxi-meters are more reliable than others in intensive care.6
After patient sensor application, a delay in obtaining dataensues while the oximeter recognises the pulse waveformand calculates HR and SpO2; both values are then displayedsimultaneously. During resuscitation, HR is of primaryimportance in determining need for intervention—for exam-ple, intubation, chest compressions. Thus prompt acquisitionof accurate HR is ideal. The sensor can be applied in a numberof ways. We sought to determine which resulted in thequickest display of accurate HR.
METHODSWe conducted a randomised crossover study of infants in ourintensive and special care nurseries. All were stable andmonitored with oximetry and electrocardiography (ECG).Measurements were taken in the supine position via a sensorapplied to the right wrist (infants ,1500 g) or palm(.1500 g). The sensor was secured using Coban wrap (3MHealth Care, St Paul, Minnesota, USA).We studied the Masimo Radical (Masimo Corporation,
Irvine, California, USA) oximeter (fig 1, A), using anaveraging interval of two seconds with maximal sensitivity.This oximeter has a patient cable (B) to which the sensor (C)is attached. The manufacturers recommend connecting thesensor to the cable and to the patient before switching theoximeter on. We wished to avoid the delay incurred switch-ing on the machine. Thus, with the machine switched on, weapplied the LNOP Neo-L sensor to each infant on threeoccasions using the following methods:
(1) sensor connected to cable, then applied to infant;
(2) sensor connected to cable, applied to investigator’s finger,then to infant;
(3) sensor applied to infant, then connected to cable.
The investigator applying the sensor and the order of thesemethods were allocated randomly. The times taken to applythe sensor, to display data, and to display accurate HR—thatis, that matched the ECG—were recorded with a stopwatch.The number of accurate first displayed HRs for each methodwas noted. Data were analysed using SPSS. Means werecompared using paired t tests.
RESULTSWe studied 40 babies of various weight (mean (SD) 1659(991) g), gestational age (29 (4.8) weeks) and postnatal age(22 (31) days).The time taken to apply the sensor using method 3
compared with method 1 was slightly longer but this was notsignificant (table 1). The time taken by the oximeter todisplay the correct HR using method 3 was significantlyshorter (mean (SD) difference 10 (20) seconds, p = 0.004)(table 1). Combining these time periods gave the time foraccurate HR data to be displayed. The quickest method was 3,followed by 2, then 1. The difference between methods 1 and3 was significant (mean (SD) difference 7 (20) seconds, p =0.047); the difference between methods 2 and 3 was not. Theproportion of accurate first displayed HRs was 80%, 28%, and93% for methods 1, 2, and 3 respectively. Thus method 3 gavethe quickest, most accurate HR data.
Abbreviations: ECG, electrocardiography; HR, heart rate; SpO2,oxygen saturation
Figure 1 Masimo Radical pulse oximeter (A) with patient cable (B) andsensor (C).
F84
www.archdischild.com
A1
DISCUSSIONThis study does not specifically address the question of whichmethod of sensor application most rapidly obtains anaccurate SpO2. We believe that HR monitoring is moreimportant than SpO2 in effectively guiding neonatal resusci-tation; thus it was the focus of this study.An oximeter switched on with the sensor connected will
immediately try to calculate data. If the sensor is not appliedto a person, the oximeter tries to interpret environmentalstimuli and may generate artificial signal. This delays thedisplay of data when the sensor is subsequently applied to aninfant (method 1). When the sensor is first applied to aninvestigator (method 2), data are acquired more quickly. Datafrom the investigator are initially averaged, however, and thisoften leads to the display of an erroneous initial HR. Thesubsequent delay in displaying the correct HR is variable(mean (SD) 9 (7) seconds here). The display of incorrect HRduring resuscitation is concerning as it may promptinappropriate intervention or inaction; thus, in the absenceof an ECG to assess accuracy of the oximeter HR, we adviseagainst using method 2.A seven second difference in obtaining data, although not
large, may be clinically important during resuscitation. Thisstudy assessed stable infants who were not being resusci-tated. It is possible that it may take longer to apply a sensorand obtain data during delivery room resuscitation. Thealgorithms used by the oximeter, however, do not change.The differences observed may thus become greater and moreclinically important.
We assessed the Masimo Radical oximeter; this studyshould be repeated for other oximeters to confirm thesuperiority of this method of sensor application.
CONCLUSIONIn intensive and special care settings, applying the sensor tothe right hand or wrist before connection to the pulseoximeter results in quicker acquisition of accurate HR data ininfants compared with other techniques. This method ofapplication should be preferred during resuscitation.
ACKNOWLEDGEMENTSWe thank Professor Neil Finer whose original observation promptedthis study. CPFO’D is supported in part by the Royal Women’sHospital Postgraduate Degree Scholarship. PGD is supported by anNHMRC Practitioner Fellowship.
Authors’ affiliations. . . . . . . . . . . . . . . . . . . . .
C P F O’Donnell, C O F Kamlin, P G Davis, C J Morley, Division ofNewborn Services, Royal Women’s Hospital, Melbourne, AustraliaC P F O’Donnell, P G Davis, University of Melbourne, Australia
Competing interests: none declared
Correspondence to: Dr O’Donnell, Research Fellow in NeonatalPaediatrics, Royal Women’s Hospital Melbourne, 132 Grattan Street,Carlton, Victoria 3053, Australia; [email protected]
Accepted 23 July 2004
REFERENCES1 International guidelines for neonatal resuscitation: an excerpt from the
guidelines 2000 for cardiopulmonary resuscitation and emergencycardiovascular care: international consensus on science. Pediatrics2000;106:e29.
2 Owen CJ, Wyllie JP. Determination of heart rate in the baby at birth.Resuscitation 2004;60:213–17.
3 Theophilopoulos DT, Burchfield DJ. Accuracy of different methods of heartrate determination during simuluated neonatal resuscitations. J Perinatol1998;18:65–7.
4 Kattwinkel J. Evaluating resuscitation practices on the basis of evidence:the findings at first glance may seem illogical. J Pediatr2003;142:221–2.
5 Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the delivery room.Indian Pediatr 2001;38:762–6.
6 Hay WW, Rodden DJ, Collins SM, et al. Reliability of conventional and newpulse oximetry in neonatal patients. J Perinatol 2002;22:360–6.
Table 1 Time taken to apply the sensor and display datafor each method of application
Method 1 Method 2 Method 3
Time to apply sensor (s) 9 (2) 10 (2) 12 (2)Time to display accurate HRfrom sensor application (s)
23 (20) 18 (17) 13 (7)
Total time to display accurateHR data (s)
32 (21) 28 (17) 25 (7)
Values are mean (SD). Method 1, sensor connected to oximeter, then toneonate; method 2, sensor connected to oximeter, applied toinvestigator’s finger, then to neonate; method 3, sensor applied toneonate, then connected to oximeter.HR, Heart rate.
Pulse oximetry in neonates F85
www.archdischild.com
A2
FEASIBILITY OF AND DELAY IN OBTAINING PULSE OXIMETRY
DURING NEONATAL RESUSCITATION
COLM P. F. O’DONNELL, MB, MRCPI, MRCPCH, C. OMAR F. KAMLIN, MB, BS, MRCPCH, PETER G. DAVIS, MD, FRACP,AND COLIN J. MORLEY, MD, FRACP, FRCPCH
Application of the sensor to newly born infants before connection to a pulse oximeter increases the reliability and speed with
which data are displayed. Data are available in most infants within 90 seconds of birth. Oximetry may be useful in guiding
interventions during resuscitation. (J Pediatr 2005;147:698-9)
Consensus statements advise that newborns’ need for and response to resuscitation be determined clinically.1 Sparked bydebate on the optimal oxygen concentration for neonatal resuscitation,2,3 a role for pulse oximetry has been suggested.4,5
It is not known how often or how quickly pulse oximetry data may be obtained from newborns anticipated to needresuscitation. Since January 2004, we have made observational studies of delivery room (DR) resuscitation at our hospital withendorsement from our Research and Ethics Committee. Techniques used included pulse oximetry. From January to March2004, we connected the sensor to the oximeter before applying the sensor to the infant. In April 2004, we found that applyingthe sensor to infants in our intensive and special care nurseries before connecting it to the oximeter yielded an accurate heartrate more quickly.6 We thus changed to this method of sensor application in the DR at this time.
The objectives of the present study were to evaluate newborns anticipated to need resuscitation in the DR to determine (1)how often pulse oximetry data are obtained, (2) the success rates and delays in obtaining data associated with different methods ofsensor application, and (3) the success rates in very low birth weight (VLBW) infants (< 1500 g) compared with heavier infants.
METHODSWhen available, 1 of 2 investigators attended deliveries andmade detailed recordings.
A digital video camera was mounted above the resuscitation cot to acquire a clear view ofthe infant. An L-NOP Neo-L sensor of a Masimo Radical (Masimo Corporation, Irvine,Calif ) pulse oximeter was placed on the infant’s right hand or wrist as soon as practicableafter delivery and secured with Coban wrap (3M Healthcare, St. Paul, Minn). The oxim-eter was set to acquire data with maximal sensitivity and to average data over 2 seconds.
To assess how often and how quickly data were obtained, we reviewed all videos andidentified those in which pulse oximetry was used. The infant’s gestational age and birthweight, as well as the method of sensor application, were noted. The times taken to applythe sensor, for the oximeter to display data, and from birth to data display were recorded.These time intervals were compared for cohorts before and after the change in the methodof sensor application and for VLBW and heavier infants. Data were analyzed using SPSSsoftware (SPSS Inc., Chicago, Ill).
RESULTSPulse oximetry was used in 115 of the 122 videos obtained. The mean (standard
deviation, range) gestational age and birth weight of these infants were 32 (6, 23 to 42)weeks and 1830 (1064, 495 to 4930) g, respectively. Of the 115 infants, 60 (52%) wereVLBW. The maximal respiratory support given to these infants was as follows: nonein 29, supplemental oxygen in 9, mask continuous positive airway pressure (CPAP) or
CPAP Continuous positive airway pressureDR Delivery room
PPV Positive-pressure ventilationVLBW Very low birth weight
698
From the Division of NewbornServices, Royal Women’s Hospital,Melbourne, Australia; Departmentsof Obstetrics and Gynaecology, Uni-versity of Melbourne, Melbourne,Australia; and Murdoch Children’s Re-search Institute, Melbourne, Australia.Supported by a RWH PostgraduateDegree Scholarship (CPFO) and aNHMRC Practitioner Fellowship(PGD).Submitted for publication Apr 15,2005; last revision received Jun 28,2005; accepted Jul 18, 2005.Reprint requests: Dr. ColmO’Donnell, Neonatal Paediatrician, Di-vision of Newborn Services, RoyalWomen’s Hospital Melbourne, 132Grattan Street, Carlton, Victoria 3053,Australia. E-mail: [email protected]/$ - see front matterCopyrightª 2005 Elsevier Inc. All rightsreserved.
10.1016/j.jpeds.2005.07.025
A3
positive-pressure ventilation (PPV) in 19, nasal CPAP in 28,and endotracheal PPV in 30. A Neopuff (Fisher and Paykel,NewZealand) delivering 100% oxygen was used for respiratorysupport. No infant received cardiac massage or medications,and all survived beyond the DR.
Oximetry data were obtained from 105 of the 115 (91%)infants. Data were always recorded if the sensor was applied tothe infant before it was connected to the oximeter (see Table).The time from sensor application to data display (ie, how longit took the machine to interpret signals) was substantiallyshorter if the sensor was applied to the infant first (see Table).Applying the sensor to the infant first produced a data displayby 92 seconds of life in 90% of infants. Data were alwaysobtained from VLBW infants irrespective of how the sensorwas applied. Response time was not affected by the use ofrespiratory support.
DISCUSSIONThe findings of the present study demonstrate that it
is possible to obtain pulse oximetry data during neonatalresuscitation. When an oximeter is switched on with a sensorconnected, it immediately attempts to generate data. If thesensor is not applied to an infant, then it will interpret envi-ronmental ‘‘noise’’ and generate an artifactual signal. Whenthe sensor is then applied, the oximeter averages this arti-factual signal with the true signal detected from the infant,leading to a delay in data display. This problem can be avoidedby first applying the sensor to the infant, because then theoximeter immediately interprets signals from the infant. Thesuperior reliability and speed in data display with this methodof sensor application suggest that it should always be used inthe DR.
We did not prospectively compare sensor applicationtechniques in a randomized fashion; thus it is possible thatthe improvement associated with the change in method ofsensor application was due to experience gained in the use ofpulse oximetry over time. If this were true, then it should be
reflected in an improvement in the time taken to apply thesensor, the human element in obtaining these data. Thistime did not differ appreciably between the 2 periods, however.
Data were always obtained from VLBW infantsirrespective of how the sensor was applied. We believe that thisreflects the difficulties sometimes encountered in opposingthe emitter and detector windows of the sensor around thewrist or palm of larger infants.
We investigated only a single pulse oximeter, and do notknow whether our results are applicable to other models. Also,although we have demonstrated the feasibility of pulse oxim-etry in the DR, the effectiveness of this technique remainsunproven. There is currently insufficient information to definenormal and abnormal oxygen saturation levels in the first fewminutes of life, or the values that should be targeted for infantsresuscitated in the DR. Until these values are defined, webelieve that pulse oximetry during neonatal resuscitation ismost useful in providing a continuous, objective display ofheart rate.
We thank Professor Neil Finer, whose observation prompted thisstudy.
REFERENCES1. International Consensus on Science. International guidelines forneonatal resuscitation: an excerpt from the 2000 guidelines for cardiopul-monary resuscitation and emergency cardiovascular care. Pediatrics 2000;106:e29.2. Saugstad OD, Ramji S, Vento M. Resuscitation of depressed newborninfantswith ambient air or pure oxygen: ameta-analysis. BiolNeonate 2005;87:27-34.3. Tan A, Schulze A, O’Donnell CP, Davis PG. Air versus oxygen forresuscitation of infants at birth. Cochrane Database Syst Rev 2005;(2):CD002273.4. Saugstad OD. Resuscitation with room air or oxygen supplementation.Clin Perinatol 1998;25:741-56.5. Kattwinkel J. Evaluating resuscitation practices on the basis of evidence:the findings at first glance may seem illogical. J Pediatr 2003;142:221-2.6. O’Donnell CPF, Kamlin COF, Davis PG, Morley CJ. Obtaining pulseoximetry data in neonates: a randomised crossover study of sensor applicationtechniques. Arch Dis Child Fetal Neonatal Ed 2005;90:F84-5.
Table. Failure rate and times for sensor application, for oximeter to generate and display data, and from birthto data display in delivery room resuscitation by method of sensor application
Sensor applied to infant afterconnection to oximeter (n = 37)
Sensor applied to infant beforeconnection to oximeter (n = 78)
Failed to display data (%) 10 (27) 0 (0)Time to apply sensor (sec)* 20 (16-24) 18 (15-22)Time to display data (sec)* 41 (20-78) 15 (11-16)Time from birth to data display (sec)* 100 (74-157) 68 (59-80)
*Data are median (interquartile range).
Feasibility Of And Delay In Obtaining Pulse Oximetry During NeonatalResuscitation 699
A4
ORIGINALARTICLES
OXYGEN SATURATION IN HEALTHY INFANTS IMMEDIATELY AFTER BIRTH
C. OMAR F. KAMLIN, COLM P.F. O’DONNELL, PETER G. DAVIS, AND COLIN J. MORLEY
Objective Because the optimal concentration of oxygen (FiO2) required for stabilization of the newly born infant has notbeen established, the FiO2 is commonly adjusted according to the infant’s oxygen saturation (SpO2). We aimed to determinethe range of pre-ductal SpO2 in the first minutes of life in healthy newborn infants.Study design We applied an oximetry sensor to the infant’s right palm or wrist of term and preterm deliveries immediatelyafter birth. Infants who received any resuscitation or supplemental oxygen were excluded. SpO2 was recorded at 60 secondintervals for at least 5 minutes and until the SpO2 was >90%.Results A total of 205 deliveries were monitored; 30 infants were excluded from the study. SpO2 readings were obtainedwithin 60 seconds of age from 92 of 175 infants (53%). The median (interquartile range) SpO2 at 1 minute was 63%(53%-68%). There was a gradual rise in SpO2 with time, with a median SpO2 at 5 minutes of 90% (79%-91%).Conclusion Many newborns have an SpO2 <90% during the first 5 minutes of life. This should be considered when choosingSpO2 targets for infants treated with supplemental oxygen in the delivery room. (J Pediatr 2006;148:585-9)
The transition from fetus to newborn is a complex physiological process. There is growing interest in the use of pulseoximetry to assess the condition of infants immediately after birth.1,2 The subcommittee of the International LiaisonCommittee on Resuscitation have noted the paucity of information on oxygen saturations (SpO2) of healthy term and
preterm infants during the first minutes of life. They have called for more data before making “evidence-based recommendationson the meaning of pulse oximetry measurements in sick babies at birth.”3 Previous reports describe term infants deliveredvaginally,4 infants who were resuscitated or received supplemental oxygen (O2),5,6 or who were monitored with technology thatis now outdated.7-9 The aim of our study was to describe the SpO2 of healthy newly born preterm and term infants using newergeneration pulse oximetry.
METHODSWe conducted a prospective observational study of SpO2 in newly born infants
between May 2004 and April 2005. The study was endorsed by the Human Research andEthics Committees of the Royal Women’s Hospital, Melbourne, Australia. Verbalparental consent for the study was obtained before delivery. The investigating team wasnot involved in the care of the infants in the delivery room.
Infants !31 weeks gestation who were not anticipated to need resuscitation werestudied when an investigator was available to attend their delivery with a pulse oximeter.A stopwatch was started when the cord was clamped. All infants were dried and wrappedwith warmed towels. Infants were excluded when they received supplemental O2 orassisted ventilation, or when we were unable to obtain oximetry data. We used theMasimo Radical (Masimo Corporation, Irvine, Calif) set to detect a signal with maximalsensitivity and averaged it over 2 seconds. The sensor was applied either to the palm of theright hand or to the right wrist to obtain preductal SpO2 and secured with Coban wrap(3M Health Care, St. Paul Minn). The sensor was then connected to the oximeter,because this leads to the fastest acquisition of data.3 The times taken to apply the sensorand to display data were noted. SpO2 was continuously monitored, with values recordedat 60-second intervals from birth for at least 5 minutes and until the SpO2 was !90%.
FiO2 Concentration of oxygenIQR Interquartile ranges
nCPAP Nasal continuous positive pressureSpO2 Oxygen saturation
See editorial, p 569 andrelated article, p 590
From the Division of Newborn Services,Royal Women’s Hospital, Melbourne, Aus-tralia; University of Melbourne, Australia;and Murdoch Children’s Research Institute,Melbourne, Australia.C.O.F.K. and C.P.F.O’D. are recipients ofthe Royal Women’s Hospital PostgraduateDegree Scholarship. P.G.D. is supported bya National Health and Medical ResearchCouncil Fellowship.Submitted for publication Aug 1, 2005; lastrevision received Dec 9, 2005; acceptedDec 20, 2005.Reprint requests: Dr Omar Kamlin, Re-search Fellow in Neonatal Paediatrics,Royal Women’s Hospital Melbourne,132Grattan St, Carlton, Victoria 3053, Austra-lia. E-mail: [email protected]/$ - see front matterCopyright © 2006 Elsevier Inc. All rightsreserved.10.1016/j.jpeds.2005.12.050
585
A5
Where necessary, the sensor remained attached after theinfant was given to the parents. The time taken to achieveSpO2 !75% and !90% was noted.
The presence and the quality of the oximetry data wereverified by the investigator using the display; thus the care-givers were not masked to the oximetry data. Apgar scoreswere assigned by the pediatricians or midwives caring for thebabies at delivery. The type of delivery, maternal analgesiaand method of anesthesia were recorded.
Statistical Analysis
Data are presented as means (SD) and analyzed with2-tailed t tests when normally distributed. Median and inter-quartile ranges (IQR) are provided and analyzed with non-parametric tests (2-tailed Mann–Whitney U test) when thedistribution of the variable was skewed. Multivariate regres-sion analysis was used to analyze potential confounding vari-ables contributing to the primary end point (time taken forSpO2 !90%). A P value of ".05 was considered to bestatistically significant. Data were analysed with SPSS soft-ware for Windows (SPSS, Chicago, Ill), version 11.5.
RESULTSA total of 205 deliveries were attended. Figure 1 shows
the characteristics of infants who were excluded (n # 30) and
infants who were included (n # 175). Technical difficultiesobtaining data occurred in 12 term infants, none of whom wasresuscitated or given supplemental oxygen. Patient character-istics and the time taken for sensor application and signaldetection are shown in Table I.
The median SpO2 values (IQR) at 1, 2, 3, 4, and 5minutes were 63% (53%-68%; n # 92), 70% (58%-78%; n# 164), 76% (64%-87%; n # 172), 81% (71%-91%; n #174), 90% (79%-91%; n # 175), respectively (Figure 2).For the whole group, the mean (SD) time to achieve SpO2!90% was 5.8 (3.2) minutes (range, 1.3-20.2 minutes).With multivariate regression, the most important determi-nants of the time to reach SpO2 !90% were gestational ageand presence of labor (R2 # 0.09 and 0.15; P ".001[analysis of variance]).
Of the 175 infants studied, 63 were admitted to thenursery. Eleven infants (all "34 weeks) were given supple-mental oxygen in the nursery; 8 of these infants were alsotreated with nasal continuous positive pressure (nCPAP). Noinfants were intubated or treated for persistent pulmonaryhypertension of the newborn. The median (IQR) age atwhich supplemental oxygen and/or nCPAP was started andduration of therapy were 0.5 (0.3-0.6) hours and 55 (4-74)hours, respectively. The median (IQR) SpO2 of these 11infants at 5 minutes in the delivery room was 78% (69%-82%)compared with a median of 91% (81%-91%) in the 164infants who were not given supplemental oxygen or nCPAP(P # .001).
Effect of Mode of Delivery, Prematurity, Analgesia,and Anesthetic on SpO2
The effect of mode of delivery and gestational age onSpO2 at 5 minutes are shown as medians (IQR) in Figure 3Aand B. The median SpO2 at 5 minutes was lower after electivecesarean section than after spontaneous vaginal delivery andvacuum extraction (Mann Whitney U test; P # .013 and .001,respectively). The median SpO2 at 5 minutes was significantlylower in preterm infants than term infants (P ".001). Thetime for each group to reach a SpO2 !75% and !90% areshown in Table II. We found no association between SpO2and maternal analgesia or anesthesia (Figure 3C and D).
DISCUSSIONThe appropriate FiO2 to use for neonatal resuscitation
is subject to debate. It has been suggested that the FiO2 couldbe determined by monitoring infants’ SpO2 by using pulseoximetry.11,12 Surveys suggest that many clinicians are alreadyusing this approach.13,14 There is limited information aboutinfants’ SpO2 in the minutes after birth. Toth et al measuredthe pre- and post-ductal SpO2 of 50 vaginally born healthyterm infants with an older generation pulse oximeter.15 Theyfound SpO2 at 2 minutes of age as low as 34%, and that theseinfants took 12 to14 minutes to reach SpO2 values !95%,and that pre-ductal SpO2 rose more quickly than the post-ductal value. Rao et al used an older oximeter to study the
Figure 1. Profile of infants enrolled.
586 Kamlin et al The Journal of Pediatrics • May 2006
A6
pre-ductal SpO2 of 95 infants who were asphyxiated andresuscitated and 30 control infants who were not resuscitat-ed.16 The resuscitated infants were part of the Resair 2study,17 a multicenter trial comparing air and 100% O2 forresuscitation of infants with birth weight !999 g who hadbradycardia and poor respiratory effort at birth. The maturityand mode of delivery of the infants in Rao’s study were notclear.18 Data were not obtained from 12 (10%) infants, and ittook longer to obtain data from infants who were asphyxiated.In the control group, SpO2 values as low as 43% were foundat 1 minute of age and rose to a mean of 90% by 15 minutes.In infants who were asphyxiated, SpO2 rose more slowly;however, the results were not reported according to the FiO2used, thus the effect of using 21% or 100% oxygen was notclear. Saugstad et al subsequently reported the SpO2 valuesfrom 229 of the 591 infants recruited to Resair 2.19 Theyreported that median SpO2 rose from about 65% at 1 minuteto 90% at 5 minutes and rose equally quickly whether infantswere resuscitated with air or 100% oxygen. The oximeter(s)used, site of sensor application, gestational age of the infants,and modes of delivery were not reported.
Using the Masimo SET technology, we measured theSpO2 of newly born infants who were not resuscitated orgiven supplemental oxygen. In addition, we examined theeffect of preterm delivery, mode of delivery, presence of labor,and maternal analgesia and anesthesia. We did not obtaindata from 12 (6%) infants. All were active term babies. Webelieve that, as has been described, poor alignment of the lightemitting diode and detector was responsible.20 The SpO2 inour infants closely resemble those found by others in non-resuscitated term infants21,22 and mildly compromised infantsresuscitated with either air or 100% oxygen.23,24
Our findings suggest that gestation and the presence oflabor have an effect on SpO2 in the minutes after birth. Likeother authors,5 we did not demonstrate an effect of eithermaternal analgesia or anesthesia on SpO2. Oximetry data canbe obtained within 1 to 2 minutes after birth and gives acontinuous, non-invasive measure of SpO2 and heart rate.Other authors have had difficulties obtaining data with olderpulse oximeters26; this may be because of low peripheral
perfusion or motion artifact. The Masimo Radical pulseoximeter, uses the same principles as conventional pulseoximetry, but also has signal processing algorithms to reduce“noise” (eg, poor signal in low perfusion states or patientmotion).
With a 2-second averaging interval, the oximetertracked the rapidly changing SpO2 during postnatal adapta-tion. We obtained data faster (median, 1.2 versus 2.3 min-utes) than an earlier study assessing this oximeter in preterminfants during resuscitation.27 This difference may be causedby different methods of sensor application, because we havealso obtained data quickly at high-risk deliveries (median, 68seconds).28
In summary, we have demonstrated the feasibility of
Figure 2. Box plots showing the median, quartiles, range, (1.5 times thequartile on that side), outliers, and extreme values for SpO2 at eachminute after birth for the first 5 minutes (N # number of patients inwhom SpO2 was obtained). A number "175 indicates that SpO2 was notobtained in all cases.
Table I. Demographic data and times taken to apply patient sensor, signal processing, and time from birthto display oximetry data by gestation
All infants(N ! 175)
Preterm(<37 weeks; N ! 54)
Term infants(>37 weeks; N ! 121)
Gestation (weeks) 37.5 (3.0) 33.5 (1.8) 39.3 (1.3)Birth weight (g) 2953 (867) 1965 (557) 3393 (560)5 minute Apgar 9 (9–9)* 9 (9–9)* 9 (9–9)*1 minute Apgar† 8 (7–9)* n # 92 8 (7–9)* n # 37 8 (8–9)* n # 55Time to apply sensor (seconds) 58 (21) 53 (16) 60 (23)Time for signal processing (seconds) 16 (4) 16 (3) 16 (5)Total time from birth to first data (seconds) 74 (22) 71 (20) 75 (23)
Data are mean (SD) or * median (IQR).Times are measured from the point of cord clamping.No statistically significant differences seen between the groups.†Of infants for whom O2 saturations available at 1 minute.
Oxygen Saturation In Healthy Infants Immediately After Birth 587
A7
Figure 3. Box plots showing the median, quartiles, range, (1.5 times the quartile on that side) outliers, and extreme values, of SpO2 at 5 minutes frombirth by: A, mode of delivery; B, maturity; C, maternal analgesia; and D, maternal anaesthesia. SVD, Spontaneous vaginal delivery; Vacuum, a vaginaldelivery assisted by vacuum extraction (Ventouse) delivery; Forceps, a vaginal delivery assisted by forceps; EM.CS, emergency cesarean section delivery;Elect.CS, elective cesarean section delivery; Preterm, delivery before 37 weeks gestation; Term, delivery from 37 weeks gestation; N2O, nitrous oxideanalgesia; Opioid, analgesia with narcotic; Spinal, delivery with spinal anesthetic; General, delivery with general anesthetic.
Table II. Comparison of times in minutes (median [IQR]) for infants to attain SpO2 >75% and >90% bymode of delivery, presence of labor and by gestational age (<37 wk versus >37 wk)
NTime to reach anSpO2 >75% (min) P*
Time to reach anSpO2 >90%(minutes) P*
Vaginal birth 68 2.4 (1.6–3.7) .006 4.0 (3.0–6.1) .001Abdominal birth 107 3.5 (2.0–4.8) 5.9 (3.9–7.9)Labor 137 2.7 (1.7–4.2) .004 4.7 (3.2–6.4) ".001Not in labor 38 4.1 (2.6–5.4) 7.0 (5.1–10.0)GA !37 weeks 121 2.5 (1.6–4.0) ".001 4.7 (3.3–6.4) ".001GA "37 weeks 54 4.2 (2.7–6.1) 6.5 (4.9–9.8)
GA, Gestational age.*Mann Whitney U test used to compare data.
588 Kamlin et al The Journal of Pediatrics • May 2006
A8
pulse oximetry in the delivery room. We have shown thatinfants !31 weeks gestation who receive neither assistedventilation nor supplemental oxygen have a gradual rise inSpO2 during the first 5 minutes of life. This period is slowerin premature infants and those born by cesarean sectionwithout labor. We did not demonstrate an effect of maternalanalgesia or anesthesia. The range of SpO2 during this timeextends well below those currently targeted in our neonatalintensive care units. Despite the similarities in the range ofSpO2 seen in our infants and those of term infants who wereasphyxiated studied by other groups,9,30 caution must beexercised before assuming that the range of SpO2 seen inhealthy infants applies to sick preterm and term in-fants.10,25,29
REFERENCES1. Kattwinkel J. Evaluating resuscitation practices on the basis ofevidence: the findings at first glance may seem illogical. J Pediatr2003;142:221-2.2. Saugstad OD. Resuscitation with room-air or oxygen supplementation.Clin Perinatol 1998;25:741-56, xi.3. AHA/AAP ILCOR 2005 Worksheet—room air/oxygen. Available at:http://www.americanheart.org/downloadable/heart/1105541784766n.RA%20vs%20O2.JG.30Nov04.Final.pdf.4. Toth B, Becker A, Seelbach-Gobel B. Oxygen saturation in healthynewborn infants immediately after birth measured by pulse oximetry. ArchGynecol Obstet 2002;266:105-7.5. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.6. Saugstad OD, Ramji S, Rootwelt T, Vento M. Response to resuscita-tion of the newborn: early prognostic variables. Acta Paediatr 2005;94:890-5.7. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.8. Saugstad OD, Ramji S, Rootwelt T, Vento M. Response to resuscita-tion of the newborn: early prognostic variables. Acta Paediatr 2005;94:890-5.9. Toth B, Becker A, Seelbach-Gobel B. Oxygen saturation in healthynewborn infants immediately after birth measured by pulse oximetry. ArchGynecol Obstet 2002;266:105-7.10. O’Donnell CP, Kamlin CO, Davis PG, Morley CJ. Obtaining pulseoximetry data in neonates: a randomised crossover study of sensor applicationtechniques. Arch Dis Child Fetal Neonatal Ed 2005;90:F84-5.11. Kattwinkel J. Evaluating resuscitation practices on the basis of evidence:the findings at first glance may seem illogical. J Pediatr 2003;142:221-2.
12. Saugstad OD. Resuscitation with room-air or oxygen supplementation.Clin Perinatol 1998;25:741-56, xi.13. O’Donnell CPF, Davis PG, Morley CJ. Use of supplementary equip-ment for resuscitation of newborn infants at tertiary perinatal centres inAustralia and New Zealand. Acta Paediatr 2005;94:1261-5.14. O’Donnell CP, Davis PG, Morley CJ. Positive pressure ventilation atneonatal resuscitation: review of equipment and international survey of prac-tice. Acta Paediatr 2004;93:583-8.15. Toth B, Becker A, Seelbach-Gobel B. Oxygen saturation in healthynewborn infants immediately after birth measured by pulse oximetry. ArchGynecol Obstet 2002;266:105-7.16. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.17. Saugstad OD, Rootwelt T, Aalen O. Resuscitation of asphyxiatednewborn infants with room air or oxygen: an international controlled trial: theResair 2 study. Pediatrics 1998;102:e1.18. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.19. Saugstad OD, Ramji S, Rootwelt T, Vento M. Response to resuscita-tion of the newborn: early prognostic variables. Acta Paediatr 2005;94:890-5.20. Barker SJ, Hyatt J, Shah NK, Kao YJ. The effect of sensor malposi-tioning on pulse oximeter accuracy during hypoxemia. Anesthesiology1993;79:248-54.21. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.22. Toth B, Becker A, Seelbach-Gobel B. Oxygen saturation in healthynewborn infants immediately after birth measured by pulse oximetry. ArchGynecol Obstet 2002;266:105-7.23. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.24. Saugstad OD, Ramji S, Rootwelt T, Vento M. Response to resuscita-tion of the newborn: early prognostic variables. Acta Paediatr 2005;94:890-5.25. House JT, Schultetus RR, Gravenstein N. Continuous neonatal eval-uation in the delivery room by pulse oximetry. J Clin Monit 1987;3:96-100.26. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.27. Kopotic RJ, Lindner W. Assessing high-risk infants in the deliveryroom with pulse oximetry. Anesth Analg 2002;94:S31-6.28. O’Donnell CP, Kamlin CO, Davis PG, Morley CJ. Feasibility of anddelay in obtaining pulse oximetry during neonatal resuscitation. J Pediatr2005;147:698-9.29. Rao R, Ramji S. Pulse oximetry in asphyxiated newborns in the deliveryroom. Indian Pediatr 2001;38:762-6.30. Saugstad OD, Ramji S, Rootwelt T, Vento M. Response to resuscita-tion of the newborn: early prognostic variables. Acta Paediatr 2005;94:890-5.
Oxygen Saturation In Healthy Infants Immediately After Birth 589
A9
Pulse oximetry and newborn infants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pulse oximetry for monitoring infants inthe delivery room: a reviewJ A Dawson, P G Davis, C P F O’Donnell, C O F Kamlin, C J Morley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
During the first few minutes of life,oxygen saturation (saturation bypulse oximetry, SpO2) increases from
intrapartum levels of 30–40%.1 In algo-rithms for neonatal resuscitation publishedby the International Liaison Committee forResuscitation,2 European ResuscitationCouncil3 and Australian ResuscitationCouncil,4 clinical assessment of an infant’scolour (a measure of oxygenation) andheart rate are used as major action points.However, studies have shown that clinicalassessment of colour during neonataltransition is unreliable.5 6 O’Donnell et al6
showed that the SpO2 at which observersperceived infants to be pink varied widely,ranging from 10% to 100%. Assessingcolour is difficult and therefore is a poorproxy for tissue oxygenation during thefirst few minutes of life.Kattwinkel7 suggested pulse oximetry
may help achieve normoxia in the deliveryroom. The American Heart Association8
suggests that ‘‘administration of a variableconcentration of oxygen guided by pulseoximetrymay improve the ability to achievenormoxia more quickly’’. Although ‘‘nor-moxia’’ and an acceptable time to achievethis during neonatal transition have notbeen defined, Leone and Finer9 advocate atarget ‘‘SpO2 of 85 to 90% by three minutesafter birth for all infants except in specialcircumstances’’—for example, diaphrag-matic hernia or cyanotic congenital heartdisease. International surveys show thatoximetry is increasingly used during neo-natal resuscitation.10 11
To date, there are no evidence-basedguidelines for using oximetry to measurean infant’s SpO2 and to guide interventionsduring neonatal transition after birth. Wereviewed the literature to evaluate theevidence on the use of SpO2 measurementsimmediately after birth.
HOW DOES PULSE OXIMETRYWORK?Pulse oximetry measures SpO2 continu-ously and non-invasively, without theneed for calibration, and correlates closelywith arterial oxygen saturation.12 Pulseoximetry is based on the red and infraredlight-absorption characteristics of oxyge-nated and deoxygenated haemoglobin. Asensor is placed around a hand or foot and
two light-emitting diodes send red andinfrared light through to a photodetectoron the other side. The changes in absorp-tion during the arterial pulsatile flow andnon-pulsatile component of the signal areanalysed. SpO2 is estimated from thetransmission of light through the pulsatiletissue bed. With each heartbeat, there is asurge of arterial blood that momentarilyincreases arterial blood volume. This resultsin more light absorption during surges. Aspeaks occur with each heartbeat, heart ratecan also be measured.
CAN SpO2 BE SUCCESSFULLYMEASURED IN THE MINUTES AFTERBIRTH?Seven studies reported between 20% and100% success in obtaining SpO2 measure-ments by 1 min after birth.5 13–18 By 5 min,the success rate improved to between 63%and 100%.13 15–17 19–23 The most commonreason for failing to obtain a measure-ment was motion artefact5 16 18–20 24; otherswere the presence of vernix,25 low perfu-sion,25 oedema,15 high ambient light,14 24
large infants,24 cracked and wrinkledskin,15 or acrocyanosis.15 Artefactsoccurred less often in more recent studieswhere Masimo signal extraction technol-ogy (SET) was used.18 25
WHERE SHOULD THE OXIMETERSENSOR BE APPLIED?In early studies, investigators placed thesensor over the right Achilles tendon,20 25–27
the forefoot19 or midfoot.22 28 Later studiesfound that measurements were obtainedfastest from the right hand,15 probablyowing to better perfusion, higher bloodpressure and oxygenation in preductalvessels.14 29 Preductal readings were signifi-cantly higher than postductal readingssoon after birth (p,0.05).5 15 22 By 17 minafter birth, there was no longer a significantdifference between preductal and postduc-tal measurements (p,0.05).5 15 22
HOW QUICKLY CAN AN SpO2
READING BE OBTAINED?A sensor can be applied to a baby within15–20 s of birth,18 21 with the first dataobtained at about 50 s after birth.15 18 21 Nostudies obtained SpO2 data on most infantsbefore 1 min after birth.14 25 26 29 When the
Masimo sensor and monitor were used,readings were obtained faster than whenthe sensor was applied to the infant beforeconnecting it to the oximeter.21
HOW DO SpO2 READINGSCHANGE IN THE FIRST FEWMINUTES AFTER BIRTH?Some studies report the range of SpO2 at1, 5 or 10 min (tables 1 and 2); othersreport the time taken to reach a prede-termined SpO2 (table 3). These studiesshow increases in SpO2 from about 60%at 1 min, but the levels vary widely, withsome infants taking .10 min to exceed90%. Therefore, it may not be appropriateto identify specific SpO2 levels at certaintimes after birth, which can be used as atrigger to alter an infant’s treatment.
DOES THE TYPE OF OXIMETERALTER THE SpO2 RESULTS?Early oximeters had motion artefact.5 16 18–
20 24 This has been improved in neweroximeters.25 33 To determine whether thenewer oximeters were more reliable thanearlier models in the delivery room,Kopotic compared the Masimo SET tothe Nellcor Oxismart, with sensors placedon each foot, in 15 newborns of,30 weeks’ gestation.25 The Masimo SETprovided data for 350 of 362 (96%) min,and the Oxismart provided data for 212 of362 min (59%; p=0.0014).25 Leone andFiner9 recommended that oximeters usedduring neonatal resuscitation shouldhave ‘‘minimal averaging time for theSpO2 values coupled with maximumsensitivity’’. The combination of thesefeatures allows rapid detection of changesin SpO2 and improved SpO2 measurementduring periods of low perfusion.34
DOES THE TYPE OF DELIVERY ALTERTHE SpO2 AFTER BIRTH?Harris et al20 found, using an early genera-tion oximeter, that SpO2 was much lowerin 44 term elective caesarean-section deliv-eries, when compared with 32 term infantsdelivered vaginally. The mean (standarderror, SEM) SpO2 at 1 min was 46% (3%)in the caesarean group and 61% (5%) in thevaginal delivery group (p,0.05), but by5 min there was no significant difference.They postulated that the difference was dueto the increased amount of lung fluid aftercaesarean section. Kamlin et al18 found that107 term infants born by elective caesareansection took on average 2 min longer toreach an SpO2 .90% than 68 infants bornby spontaneous vaginal delivery. Rabi et al29
found a similar difference in a cohort of 115infants. In infants of.34 weeks’ gestation,the median (interquartile range, IQR) forvaginal births at 5 min was 87% (80–95%)and that for caesarean delivery was 81%(75–83%).29 In contrast, others found no
F4 LEADING ARTICLES
www.archdischild.com
A10
significant difference in SpO2 measure-ments in infants delivered vaginally or bycaesarean section, regardless of the pre-sence or type of anaesthesia.5 14 24
DOES RESUSCITATION WITH AIROR OXYGEN AFFECT SpO2 AFTERBIRTH?Table 4 summarises trials comparingSpO2 measurements at 1, 3 and 5 minin infants with asphyxia randomised toreceive air or 100% oxygen during resus-citation. In the Resair 2 study, whichenrolled infants weighing .999 g withapnoea and bradycardia at birth, therewere no significant differences in time toreach an SpO2 of 75%. The median (95%confidence interval) time to reach anSpO2 of 75% was 1.5 (1.4 to 1.6) min in
the group receiving air versus 2.5 (1.9 to3.1) min in the group receiving oxygen(p=0.27).32 In this study, the resuscita-tors were aware of the gas used, whereasVento et al31 blinded resuscitators to thetype of gas used to resuscitate infantswith asphyxia. He found no significantdifference in time to reach an SpO2 .90%between the two groups with asphyxia.The striking result of these studies is thatresuscitating with air or 100% oxygen hadlittle effect on the change in SpO2 in thefirst 10 min after birth.
DOES ALTITUDE AFFECT SpO2
AFTER BIRTH?Gonzales and Salirrosas28 showed thatSpO2 was significantly higher in infantsborn at sea level (Lima 150 m) than in
infants born at a higher altitude (Cerro dePasco 4340 m) from 1 min to 24 h afterbirth (p,0.01).
DOES GESTATION AFFECT SpO2?There are two reports of SpO2 measure-ments in premature infants after birth. InKopotic and Lindner’s25 study of 15 infantsborn at 24–29 weeks’ gestation, the SpO2
was >80% by 4.4 (1.9–40) min (median(range)). In the study by Kamlin et al18 oninfants not receiving resuscitation, the timeto reach an SpO2 .90% was significantlylonger in 54 preterm infants at 6.5 (4.9–9.8) min (median (IQR)) than in 121 terminfants at 4.7 (3.3–6.4) min (median(IQR)) (p,0.001). Other studies includingpremature infants did not report SpO2 fordifferent gestational ages.14 32
Table 1 Observational studies measuring SpO2 in the first few minutes of life in the delivery room where no infant receivedsupplemental oxygen
StudyGestation(weeks)
Type ofoximeter
Sensorlocation n
SpO2 (%)
Comments1 min 5 min 10 min
Harris et al20 .37 Nellcor N-100 Postductal 32 61 (5)* NA NA Vaginal delivery44 46 (3)* C/S
Toth et al22 >35 Nellcor N-300 Preductal 50 NA 84 (48–99)! 92 (65–99)! 48 spontaneous deliveries,2 vacuum extractionPostductal 78 (42–97)! 89 (62–99)!
Rabi et al29 >35 Masimo Radical Preductal 45 NA 87 (80–95)` NA Vaginal delivery Calgary(1049 m)
81 (75–83)` C/S Calgary(1049 m)
Kamlin et al18 >31 Masimo Radical Preductal 175 63 (53–68)` 90 (79–91)` NA 51 preterm124 term infants
Gonzales andSalirrosas28
.37 Nellcor N-20 Postductal 37 42 (2)* NA NA Cerro de Pasco (4340 m)131 61 (1)* Lima (150 m) sea level
Gungor et al30 .37 Air-ShieldsVickers 19040
Preductal 70 69 (0.7)1 90 (2)1 NA No suction70 70 (0.7)1 80 (2)1 92 (0.4)1 Suction
C/S, caesarean section; IQR, interquartile range; NA, not available; SpO2, saturation by pulse oximetry.*Mean (SEM); !mean (range); `median (IQR); 1mean(SD).
Table 2 Observational studies measuring SpO2 in the first few minutes of life in the delivery room where some infants were treatedwith 100% oxygen
Study Gestation Type of oximeter Sensor location Resuscitation n
SpO2 %
Comments1 min 5 min 10 min
Sendaket al26
Term andpreterm
Nellcor N-100 Postductal No infant received oxygen 25 63* 89* NA Vaginal delivery34 27 (4)! 72* C/S
100% oxygen given 27 48 (5)! 69* C/S and oxygen
Houseet al14
Term andpreterm
Nellcor N-100 orOhmedaBiox 3700
Preductal 19/38 vaginal deliveries and53/62 C/S received100% oxygen
100 58 (22)` 82 (14)` 89 (6)`
No infant received oxygen 28 78 (9)` 84 (14)` 90 (5)` C/S and vaginal deliveries
Deckardtet al19
.37 weeks Nellcor N-100 Postductal 12 infants received CPAP with100% oxygen
35 40–75 range NA NA All vaginal deliveries
Porter et al24 Term Ohmeda Biox3700
Preductalor postductal
100% oxygen if poor respiratoryeffort, central cyanosis, orheart rate ,100. Number receivingoxygen not indicated
96 77 (11)` 84 (7)` 89 (6)` C/S and vaginal deliveries
Raoand Ramji16
Term andpreterm
Novametrix515A
Preductal Infants with asphyxia randomisedto receive air or 100% oxygenduring resuscitation
95 45 (20)` 84 (14)` 91 (10)` Infants with asphyxia enrolledin the Resair 2 study
Not reported 30 70 (16)` 89 (9)` 94 (2)` Non-asphyxiated controls
Dimichet al5
Not reported OhmedaBiox 3700
Preductal 7 infants received 100%supplemental oxygen
100 72 (6)` 83 (4)` 91 (5)` 63 vaginal deliveries37 C/SPostductal 63 (4)` 77 (4)` 87 (6)`
CPAP, continuous positive airway pressure; C/S, caesarean section; NA, not available; SpO2, saturation by pulse oximetry.*Mean; !mean (SEM); `mean (SD).
LEADING ARTICLES F5
www.archdischild.com
A11
CAN OXIMETRY BE USED TOMEASURE THE EFFECT OFRESUSCITATION PRACTICES?Oximetry has been used to measure theeffect of clinical interventions, such asoropharyngeal suction and endotrachealintubation during neonatal transition.Three controlled studies show that suc-tioning does seem to have a negativeeffect on oxygenation.13 23 30 O’Donnellet al36 measured the effects of attemptedendotracheal intubation on SpO2 in thedelivery room, and SpO2 often fell duringintubation attempts.
COULD OXIMETRY BE USED IN THEDELIVERY ROOM TO IMPROVEOUTCOMES?There are two studies that evaluate theuse of oximetry to guide interventionsduring neonatal transition. Deckardtet al19 used SpO2 readings at 5 min afterbirth to determine whether infantsshould receive continuous positive airwaypressure (CPAP) with a mask and 100%oxygen. CPAP was used only if the SpO2
was ,80% at 5 min and stopped once the
SpO2 reached 90%. Kopotic and Lindner25
studied 50 infants at risk for respiratoryfailure; 25 infants were managed withoutoximetry and compared with 25 managedwith oximetry. Infants managed withoximetry were less likely to be admittedto the special care nursery (32% v 52%;p=0.04). They also observed the effect ofoximetry during resuscitation in 15infants of ,30 weeks’ gestation. Initialrespiratory care was based on the infant’sclinical state and oximetry measure-ments. Oxygen was started at 100% andadjusted to achieve an SpO2 between 80%and 92%.25 The authors claim that byusing pulse oximetry they were able toreduce the fraction of inspired oxygen(FiO2) from 1.0 to, on average, 0.40. Thestudies by Kopotic and Deckardt,although non-blinded and non–randomised, suggest that oximetry canimprove short-term outcomes—forexample, admission to nursery, the useof oxygen or CPAP. We could find noreports on whether the use of SpO2
measurements immediately after birthalters long-term outcomes.
CONCLUSIONBefore oximetry is advocated for routineuse in the delivery room, more research isneeded to define normoxia, and moreimportantly, how to interpret and applySpO2 readings to clinical practice to improveshort-term and long-term outcomes.
Arch Dis Child Fetal Neonatal Ed2007;92:F4–F7.doi: 10.1136/adc.2006.102749
Authors’ affiliations. . . . . . . . . . . . . . . . . . . . . . .
J A Dawson, P G Davis, C P F O’Donnell,C O F Kamlin, C J Morley, Division of NeonatalServices, The Royal Women’s Hospital,Melbourne, Victoria, Australia
Correspondence to: J Dawson, Division ofNeonatal Services, Royal Women’s Hospital,Melbourne, 132 Grattan Street, Carlton, Victoria3053, Australia; [email protected]
Accepted 20 July 2006
Supported in part by a RWH Postgraduatedegree scholarship (CPFOD and COFK) and aNHMRC Practitioner Fellowship (PGD).
Competing interests: None.
Table 3 Time (in min) to reach specified SpO2 levels after birth
StudyGestation(weeks)
Type ofoximeter Sensor locationn Resuscitation
SpO2
Comments.75% .80% .90% 95%
Toth et al22 >35 Nellcor N-3000
Preductal 50 No infant receivedoxygen
NA NA NA 12 (2–55)* 48 spontaneousdeliveries, 2 vacuumextraction
Postductal 50 14 (3–55)*
Kamlin et al18 >31 MasimoRadical
Preductal 175 No infant receivedoxygen
NA NA 5.8 (3.2)!Range 1.3–20.2
NA 51 preterm124 term infants
Kopotic and Lindner25 , 30 MasimoRadical
Preductal 15 Infants initially received100% oxygen thenoxygen adjustedaccording to SpO2
measurements
NA 4.4(1.9–40)`
NA NA
Vento et al31 . 37 Notreported
Notreported
22 Control group NA NA 0.9 (0.4)! NA Non-asphyxiated
55 Air 2.0 (0.7)! Infants with asphyxia52 100% oxygen 1.8 (0.9)!
Rao and Ramji16 >31 Novametrix515A
Preductal 95 Infants with asphyxiarandomised to receiveair or 100% oxygenduring resuscitation
5 (3–7.25)1 NA 7.3 (4.5–11)1 NA Infants with asphyxiaenrolled in the Resair2 study
30 Not reported 2.8 (1.6–4.4)1 NA 4.3 (2.9–5.8)1 NA Non-asphyxiatedcontrol group
Saugstad et al32 >31 Notreported
Notreported
103 Air 1.5 (1.4–1.6)1 NA NA NA Infants with asphyxiaenrolled in the Resair2 study
109 100% oxygen 2.5 (1.9–3.1)1
NA, not applicable, SpO2, saturation by pulse oximetry.*Mean (range); !mean (SD); `median (range); 1median (95% CI).
Table 4 SpO2 measurements in infants with asphyxia randomised to resuscitation with air or 100% oxygen
1 min 5 min 10 min
Air Oxygen Air Oxygen Air Oxygen
Saugstad et al32 65 (11)* 61 (14)* 86 (10)* 88 (10)* 90 (6)* 91 (7)*Saugstad et al35 68 (40–82)! 63 (40–82)! 90 (66–95)! 90 (73–96)! 90 (83–96)! 92 (79–97)!
*Mean (SD); !median (5th–95th centile).
F6 LEADING ARTICLES
www.archdischild.com
A12
REFERENCES1 East CE, Colditz PB, Begg LM, et al. Update on
intrapartum fetal pulse oximetry. Aust N Z J ObstetGynaecol 2002;42:119–24.
2 International Liaison Committee on Resuscitation.Part 7: neonatal resuscitation. Resuscitation2005;67:293–303.
3 Biarent D, Bingham R, Richmond S, et al. EuropeanResuscitation Council Guidelines for Resuscitation.Resuscitation 2005;67S1:S97–133.
4 Australian Resuscitation Council. Guideline 13.1,Introduction to resuscitation of the newborn infant.2006. http://www.resus.org.au(accessed 10 Aug2006).
5 Dimich I, Singh PP, Adell A, et al. Evaluation ofoxygen saturation monitoring by pulse oximetry inneonates in the delivery system. Can J Anaesth1991;38:985–8.
6 O’Donnell CPF, Kamlin COF, Davis PG, et al.Clinical assessment of colour at neonatalresuscitation: The Mullet Study. Washington, DC:PAS, 2005.
7 Kattwinkel J. Evaluating resuscitation practiceson the basis of evidence: the findings at firstglance may seem illogical. J Pediatr2003;142:221–2.
8 American Heart Association. American HeartAssociation (AHA) Guidelines for cardiopulmonaryresuscitation (CPR) and emergency cardiovascularcare (ECC) of pediatric and neonatal patients:neonatal resuscitation guidelines. Pediatrics2006;117:1029–38.
9 Leone TA, Finer NN. Neonatal resuscitation:beyond the basics. NeoReviews 2005;6:e177–83.
10 Leone TA, Rich W, Finer NN. A survey of deliveryroom resuscitation practices in the United States.Pediatrics 2006;117:164–75.
11 O’Donnell CPF, Davis PG, Morley CJ. Use ofsupplementary equipment for resuscitation ofnewborn infants at tertiary perinatal centres inAustralia and New Zealand. Acta Paediatr2005;94:1261–5.
12 Hay WW Jr, Brockway JM, Eyzaguirre M.Neonatal pulse oximetry: accuracy and reliability.Pediatrics 1989;83:717–22.
13 CarrascoM, Martell M, Estol PC. Oronasopharyngealsuction at birth: effects on arterial oxygen saturation.J Pediatr 1997;130:832–4.
14 House JT, Schultetus RR, Gravenstein N. Continuousneonatal evaluation in the delivery room by pulseoximetry. J Clin Monit 1987;3:96–100.
15 Meier-Stauss P, Bucher HU, Hurlimann R, et al.Pulse oximetry used for documenting oxygensaturation and right-to-left shunting immediatelyafter birth. Eur J Pediatr 1990;149:851–5.
16 Rao R, Ramji S. Pulse oximetry in asphyxiatednewborns in the delivery room. Indian Pediatr2001;38:762–6.
17 Rao R, Yax S, Rao S. The role of oximetry in the first10 minutes of age after birth. Washington, DC:PAS, 2005.
18 Kamlin COF, O’Donnell CPF, Davis PG, et al.Oxygen saturation in healthy infants immediatelyafter birth. J Pediatr 2006;148:585–9.
19 Deckardt R, Schneider KT, Graeff H. Monitoringarterial oxygen saturation in the neonate. J PerinatMed 1987;15:357–60.
20 Harris AP, Sendak MJ, Donham RT. Changes inarterial oxygen saturation immediately after birth inthe human neonate. J Pediatr 1986;109:117–19.
21 O’Donnell CPF, Kamlin COF, Davis PG, et al.Feasibility of and delay in obtaining pulse oximetryduring neonatal resuscitation. J Pediatr2005;147:698–9.
22 Toth B, Becker A, Seelbach-Gobel B. Oxygensaturation in healthy newborn infants immediatelyafter birth measured by pulse oximetry. ArchGynecol Obstet 2002;266:105–7.
23 Waltman PA, Brewer JM, Rogers BP, et al. Buildingevidence for practice: a pilot study of newborn bulbsuctioning at birth. J Midwifery Womens Health2004;49:32–8.
24 Porter KB, Golhamer R, Mankad A, et al.Evaluation of arterial oxygen saturation in pregnantpatients and their newborns. Obstet Gynecol1988;71:354–7.
25 Kopotic RJ, Lindner W. Assessing high-risk infantsin the delivery room with pulse oximetry. AnaesthAnalg 2002;94:S31–6.
26 Sendak MJ, Harris AP, Donham RT. Pulse oximetryin newborn infants in the delivery room.Anaesthesiology 1985;63:433.
27 Maxwell LG, Harris AP, Sendak MJ, et al.Monitoring the resuscitation of preterm infants in thedelivery room using pulse oximetry. Clin Pediatr(Philadelphia) 1987;26:18–20.
28 Gonzales GF, Salirrosas A. Arterial oxygensaturation in healthy newborns delivered at term inCerro de Pasco (4340 m) and Lima (150 m).Reprod Biol Endocrinol 2005;3:46.
29 Rabi Y, Yee W, Chen SY, et al. Oxygen saturationtrends immediately after birth. J Pediatr2006;148:590–4.
30 Gungor S, Teksoz E, Ceyhan T, et al.Oronasopharyngeal suction versus no suction innormal, term and vaginally born infants: aprospective randomised controlled trial. AustNZ J Obstet Gynaecol 2005;45:453–6.
31 Vento M, Asensi M, Sastre J, et al. Oxidative stressin asphyxiated term infants resuscitated with 100%oxygen. J Pediatr 2003;142:240–6.
32 Saugstad OD, Rootwelt T, Aalen O. Resuscitation ofasphyxiated newborn infants with room air oroxygen: an international controlled trial: the Resair2 study. Pediatrics 1998;102:e1.
33 Bohnhorst B, Corinna PS, Poets CF. Pulseoximeters’ reliability in detecting hypoxemia andbradycardia: comparison between a conventionaland two new generation oximeters. Crit Care Med2000;28:1565–8.
34 Masimo Corporation. Radical signal extractionpulse oximeter operator’s manual. Irvine, CA:Masimo, 2004.
35 Saugstad OD, Siddarth R, Rootwelt T, et al. Responseto resuscitation of the newborn: early prognosticvariables. Acta Paediatr 2005;94:890–5.
36 O’Donnell CPF, Kamlin CO, Davis PG, et al.Endotracheal intubation attempts during neonatalresuscitation: success rates, duration, and adverseeffects. Pediatrics 2006;117:e16–21.
Controversy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Neonatal anthropometric charts: whatthey are, what they are notE Bertino, S Milani, C Fabris, M De Curtis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Over 40 years have elapsed sinceLubchenco et al1 proposed ananthropometric classification of
neonates based on the so-called intrau-terine growth charts—that is, birthweight-for-gestational age charts.
ARE NEONATALANTHROPOMETRIC CHARTSINTRAUTERINE GROWTH CHARTS?The use of charts, such as those given byLubchenco et al,1 based on the distributionof measurements taken on neonates withdifferent gestational age, should berestricted to the auxological assessmentof babies at birth. These charts, now
called neonatal anthropometric charts,must not be confused with the intrauter-ine growth charts, which are a tool formonitoring fetal growth, based on ultra-sound measurements of anthropometrictraits during pregnancy: preterm birthsare abnormal events and preterm neo-nates cannot be equated to fetuses of thesame gestational age who will be born atterm.2 When fetal growth studies arelongitudinal, both distance and velocityintrauterine growth charts may betraced.3 4 Strictly speaking, only chartsderived from longitudinal studies shouldbe called growth charts, growth being aprocess extended over time.
DOES ‘‘SMALL-FOR-GESTATIONALAGE’’ MEAN ‘‘ INTRAUTERINEGROWTH RESTRICTED’’?The terms SGA and intrauterine growthrestriction (IUGR) are often used assynonyms, although they reflect twodifferent concepts. SGA refers to a statis-tical definition, based on an auxologicalcross-sectional evaluation (prenatal orneonatal), and denotes a fetus or aneonate whose anthropometric variables(usually weight) are lower than a giventhreshold value computed on a set ofinfants having the same gestational age.SGA includes infants who have notachieved their own growth potential,because of maternal, uterine, placentaland fetal factors,5 6 as well as small butotherwise healthy infants. IUGR refers toa clinical and functional condition anddenotes fetuses unable to achieve theirown growth potential: a fetus with IUGRwould have been larger, without adverseenvironmental or genetic factors affectinggrowth. Such a condition can be assessedby ultrasonography during pregnancyby a longitudinal evaluation of fetalgrowth rate. A neonate identified asSGA by neonatal anthropometric chartsis not necessarily a case of IUGR and,
LEADING ARTICLES F7
www.archdischild.com
A13
conversely, a neonate identified as havingIUGR during the fetal period by intrau-terine growth charts may not be SGA. Thecurrent gold standard in neonatal aux-ological evaluation is based on informa-tion obtained from both neonatalanthropometric charts and intrauterinegrowth charts. Furthermore, Dopplervelocimetry can detect altered flow statesin the fetal–placental and uterine–placen-tal circulation, and may contribute to thedifferentiation between a fetus withIUGR and a fetus who is constitutionallySGA.7 8 When the prenatal growth patternis unknown, SGA may be regarded as aproxy of IUGR. An alternative proxy isbased on the prediction of birth weightbased on early ultrasound assessments offetal growth9: a negative differencebetween actual and predicted birthweight denotes IUGR. So far, there isinsufficient evidence that this alternativemethod performs better than those basedon fetal or neonatal charts.10
WHAT ABOUT RELIABILITY OFANTHROPOMETRIC ANDGESTATIONAL AGEEVALUATIONS?Weight, length and head circumference atbirth are indicators of the quality andquantity of growth: these variables mustbe evaluated using standardised instru-ments and following the techniquesrequired for accurate measurements asdescribed by Cameron.11
The validity of neonatal charts is alsobased on reliable estimates of gestationalage, expressed as complete weeks, inaccordance with international recom-mendations.12 Early ultrasound assess-ment has improved the accuracy ofestimation of gestational age,5 and thereis unanimous agreement that the bestestimation is obtained by a combinationof anamnestic—that is, based on reportedlast menstrual period—and early ultra-sound assessment.13 The a priori exclu-sion of neonates with unreliablegestational age seems more sensible thanthe a posteriori use of any statisticalmethod for detecting biologically implau-sible birth weight–gestational agepairs.12 14
SHOULD A NEONATAL CHART BE AREFERENCE OR A STANDARD?The target population is the population onwhich the chart is built and to which thechart will apply. A target population isdefined by its inclusion criteria—that is,geographical area, ethnic group, sex,single birth, live birth and so on. In theabsence of exclusion criteria regardingrisk factors for fetal growth, a chart basedon such a population is a reference, which
describes ‘‘how growth actually is’’ inthat population. Centers for DiseaseControl and Prevention growth chartsfor the US15 are a reference in the sensethat they are explicitly descriptive,although the authors recognise that somecompromises were made on developing atrue reference.16 The two anthropometriccharts elaborated by the Italian Society ofNeonatology,17 18 as well as most neonatalcharts in use, are essentially descriptivereferences. Differences between referencecharts reflect the different anthropo-metric characteristics of healthy neonatesbelonging to different populations andalso the different prevalences of riskfactors for prenatal growth in thosepopulations. For this reason, by meansof reference charts, the differences in thehealth conditions of two populations, orof one population over time, may beevaluated. On the other hand, the clinicaluse of a reference raises some methodo-logical problems, as a neonate is com-pared with a group of peers, alsoincluding infants who may have hadprenatal growth impairment; therefore,a reference might possess low sensitivityin detecting a neonate with growthanomalies. From a practical viewpoint,when the chart is based on a populationwith low prevalence of risk factors (suchas the populations of developed coun-tries), the clinical use of a reference canbe safely accepted.To avoid the methodological weakness
of clinical use of a reference, a set ofexclusion criteria can be defined, con-cerning mothers for example, hyperten-sion, diabetes or renal diseases, fetuses(genetic disorders or congenital anoma-lies), or uterine or placental factors.Highly restrictive criteria aiming toexclude all neonates exposed to anyknown risk factor for intrauterine growthdefine the characteristics of infants whofully expressed their growth potential.Such characteristics constitute a model towhich a neonate should conform, and abasis for a prescriptive standard or normthat indicates how growth should be.16
However, there is no agreement on whichdiseases should be taken into considera-tion, and some of these may even passunnoticed at birth. Moreover, it is rare tofind neonates without IUGR with lowgestational age when highly restrictiveexclusion criteria are adopted, so that anorm for a severely preterm neonate maybe difficult to draw. An example ofneonatal standards are the Italian chartsbased on a multicentre survey carried outbetween 1973 and 1979.19 Although thesecharts are the result of a noteworthy (forthat time) methodological effort, theyoverestimate the value of anthropometrictraits at low gestational age, where there
is a higher probability of including infantswith a true gestational age value abovethat assessed (at that time, ultrasoundassessment of gestational age was notcommon obstetric practice). Even if anaccurate neonatal standard were avail-able, its clinical use could be question-able: a large proportion of severelypreterm neonates have IUGR a priori,20
and are expected to be classified as SGAon the basis of such standards. Bycontrast, the use of a reference, includingneonates with different degrees of IUGR,enables the detection of preterm neonateshaving severe IUGR.
MANY LOCAL REFERENCE CHARTSOR A UNIQUE STANDARD?A much-debated topic is whether agrowth chart should be local, national orinternational. Strictly speaking, as areference chart describes the anthropo-metry of a given population, we need asmany reference charts as the number ofdifferent populations, no matter whethertheir anthropometric differences areascribable to ethnic characteristics or toenvironmental, nutritional, socioeco-nomic and health conditions.Do we really need, however, as many
standards as the number of differentpopulations? If the main reason for thedifferences emerging by comparisonbetween different reference charts is theinequality in health between poor andrich populations, these differences tend tovanish when the restrictive exclusioncriteria that define a standard populationare adopted. In this case, only onestandard could apply to all populations.The new World Health Organization childgrowth standards are based on such anassumption.21 Even full-term single-bornhealthy infants of non-smoking mothersfrom a favourable socioeconomic statusshow a residual difference in size at birthcorrelated with ethnicity—for example,1.4 cm in birth length betweenNorwegian and Indian neonates.22 Aunique standard is the right or the wrongchoice depending on whether such differ-ences are regarded as negligible or not.The extent to which the anthropometricdifferences between ethnic groups are theresult of health, socioeconomic and envir-onmental factors is still debated.23
As asserted by Karlberg et al,24 cliniciansseem to prefer local references whencommunicating with patients and theirparents, and do not seem to take seriouslyany attempt to establish an internationalstandard. Severely preterm neonates whomatch the requirements for a standardcan hardly be found; thus, neonatalcharts can be based only on a local ornational reference population.
F8 LEADING ARTICLES
www.archdischild.com
A14
TRADITIONAL POPULATION-BASED OR CUSTOMISED CHARTS?Establishing neonatal charts adjusted forfactors permanently bound to differencesin fetal growth such as sex, and single ormultiple pregnancy25 26 is indeed useful:such factors are generally taken intoaccount to trace population-based charts.The adjustment for other covariates (theso-called customising features, such asmaternal height, weight, and even mater-nal birth weight and birth weight ofprevious siblings) is gaining increasingpopularity.27–30 From a systematic reviewof the evidence, it seems that customisedcharts could be suitable to improve thedetection of IUGR.31 Nevertheless, custo-mising features reflect constitutional fac-tors but are also surrogates for acombination of parameters related to themother, such as socioeconomic level andnutrition10 30: the available data do notpermit confident inferences regarding theextent to which they induce physiologicalor pathological variations in fetalgrowth.12
HOW TO CHOOSE A CUT-OFFPOINT TO DEFINE SGANEONATES?A clinically useful threshold value woulddiscriminate neonates with IUGR, whoare at high risk of short-term and long-term growth impairment, disease anddeath, from neonates without IUGR,who are at low risk. On inspection ofneonatal morbidity and mortality ‘‘riskmaps’’—that is, a kind of geographicalmap where prefixed levels of risk areplotted as contours as a function ofgestational age (longitude) and birthweight (latitude)—it seems that neonatalrisk increases with the decrease in birthweight and gestational age.25 32 SGAneonates have long-term risk of auxolo-gical deficit,6 33 neurocognitive impair-ment,34 metabolic disorders andcardiovascular diseases.33 35 These obser-vations justify the use of neonatal charts,
but are of no help in identifying valuesthat best discriminate between infants athigh and low risk. An alternative is toadopt statistical definitions instead ofclinical ones, although the thresholdsbased on statistical criteria are onlyindirectly related to risk. In accordancewith the statistical criterion, a neonate isdefined to be SGA when his or her weightis below the 10th, 5th or 3rd centile of theneonatal chart or, under assumption of agaussian distribution, 1.5 or 2 standarddeviations below the average (whichcorrespond to the centiles 6.7 and 2.3).5
The choice of a threshold affects bothsensitivity (proportion of SGA neonatesamong those with IUGR) and specificity(proportion of AGA neonates amongthose without IUGR): the use of the 3rdcentile instead of the 10th centileincreases specificity but decreases sensi-tivity. In the case of a standard based onlyon neonates without IUGR, setting thecut-off point at the 10th centile is thesame as setting the false-positive ratio at10%—that is, a specificity of 90%. In thecase of a reference, the false-positive ratiois expected to be ,10%, as the referenceset also includes neonates with IUGR. Nounivocal criterion states that one thresh-old is better than another, and a generalagreement on the centiles to be adoptedas cut-off points would be desirable.
SHOULD NEONATAL CHARTS BEUPDATED?As regards paediatric age range, anthro-pometric charts should be updated every5–10 or 15–20 years, in conformity withthe intensity of the ‘‘secular trend ofgrowth’’ in the population.24 36 In the past25 years, developed countries haveexperienced a secular trend also in birthweight.37 Thus, more frequent updating ofneonatal charts has become necessary asa result of changes not only in parity andmaternal age and size but also in socio-economic or environmental conditions,and obstetric or neonatal care.
WHAT MODELS ARE USED TOTRACE NEONATAL CHARTS?By definition, neonatal charts are basedon data from cross-sectional studies:thus, raw non-parametric centiles of thedistribution of an auxometric variableconditional on gestational age show anuneven pattern when they are plottedversus age. The need to trace smoothcentiles derives from the assumption thatsomatic growth is a continuous process,at least at a macroscopic level, andpattern irregularity is interpreted not asthe expression of an underlying biologicalphenomenon but rather as a combinedeffect of measurement error and samplingvariability. To trace smooth growthcharts, Healy et al38 introduced a class oflinear models (Healy Rasbash Yangmethod), where the value of a givencentile at a given age is expressed aspolynomial function of age and z scorecorresponding to the centile—for exam-ple, the z score for the 3rd centile is21.88. As an alternative, Cole39 proposedthe LMS method. This sums up the age-dependent changes in the distribution ofan auxometric variable by means of threecurves that represent the degree of skew-ness (L(t)), the median (M(t)) and thecoefficient of variation (S(t)) at each age(t). This method permits the use of the zscore even in the case of non-gaussianvariables.
CONCLUSIONThe neonatal charts currently in uselargely differ as regards inclusion andexclusion criteria, techniques and instru-ments for measurement, accuracy ofassessment of gestational age and meth-ods to compute centiles. Table 1 listsseveral characteristics that a reliableneonatal chart should possess.Neonatal charts traced according to the
recommendations mentioned in table 1are of both epidemiological and clinicaluse. From an epidemiological viewpoint,a reference neonatal chart provides a
Table 1 Characteristics that a reliable neonatal chart should possess to be of both epidemiological and clinical use
Type of survey Pre-planned multicentre ad hoc studyType of chart Descriptive reference rather than an ideal prescriptive standardExclusion criteria Stillbirth, major congenital anomaliesTarget population Mono-ethnic population living in a given country at a given timeSubpopulations Females or males, single or multiple pregnancies, primiparae or multiparaeAssessment of GA Last menstrual period confirmed by early ultrasound assessmentRange of GA From 42 to 24 weeks or less, to cope with the increasing number of neonates with low GAMeasurements Use of standardised instruments and measurement techniquesTechnique to trace charts HRY method38 or LMS method39
Sample size Critical sample size concerns the more external (eg, the 3rd and 97th) centiles at lower GA, therefore, attentionshould focus on the number of severely preterm neonates, who are more difficult to recruit. Simulation indicatesthat if 100 neonates are available at 24 weeks, 95% of the HRY or LMS estimates of the 3rd centile are includedbetween centiles 1.3 and 6.3. This range narrows rapidly when GA increases (eg, at 26 weeks is betweencentiles 2.1 and 4.2) in the case that 100 neonates are sampled at each GA. Several neonates at term have pooreffect on the precision of estimates at low GA.
HRY, Healy Rasbash Yang; GA, gestational age; LMS, lambda (skewness coefficient), mu (median), sigma (coefficient of variation).
LEADING ARTICLES F9
www.archdischild.com
A15
picture of the health status of a popula-tion. The comparison of charts referringto different and clearly defined popula-tions living in the same country or indifferent countries, or to the same popu-lation in different periods, is a way ofmeasuring the extent of inequalities inhealth between populations or to monitortrends over time in response to publichealth policies.From a clinical viewpoint, a neonatal
chart is essentially a tool to detectneonates at higher risk of neonatal andpostnatal morbidity and growth impair-ment, and to compare neonatal anthro-pometric conditions with those observedduring postnatal growth. A comprehen-sive auxological evaluation of the neonateshould consider not only weight, lengthand head circumference at birth but alsofetal ultrasound biometry and Dopplervelocimetry. At present, further clinicalstudies are needed to reach a consensuson how to combine neonatal and prenatalinformation to discriminate neonateswith IUGR from those without IUGR.
Arch Dis Child Fetal Neonatal Ed2007;92:F7–F10.doi: 10.1136/adc.2006.096214
Authors’ affiliations. . . . . . . . . . . . . . . . . . . . . . .
E Bertino, C Fabris, Cattedra di Neonatologia,Dipartimento di Scienze Pediatriche, Universitadi Torino, Turin, ItalyS Milani, Istituto di Statistica Medica eBiometria, Universita di Milano, Milan, ItalyM De Curtis, Dipartimento di ScienzeGinecologiche, Perinatologia e Puericultura,Universita La Sapienza, Rome, Italy
Correspondence to: M De Curtis, Dipartimento diScienze Ginecologiche, Policlinico Umberto I, Viadel Policlinico 155, 00161 Rome, Italy;[email protected]
Accepted 12 June 2006
Competing interests: None declared.
REFERENCES1 Lubchenco LO, Hansman C, Dressler M, et al.
Intrauterine growth as estimated from liveborn birthweight data at 24 to 42 weeks of gestation.Pediatrics 1963;32:793–800.
2 Zaw W, Gagnon R, da Silva O. The risk of adverseneonatal outcome among preterm small for
gestational age infants according to neonatal versusfetal growth standards. Pediatrics2003;111:1273–7.
3 Bertino E, Di Battista E, Bossi A, et al. Fetal growthvelocity: kinetic, clinical, and biological aspects. ArchDis Child Fetal Neonatal Ed, 1996;74:F10–15.
4 Di Battista E, Bertino E, Benso L, et al. Longitudinaldistance standards of fetal growth. Acta ObstetGynaecol Scand 2000;69:165–73.
5 Bertino E, Coscia A, Tafi L, et al. Prenatal andneonatal growth. In: Nicoletti I, Benso L, Gilli G, eds.Physiological and pathological auxology. Firenze:Edizioni Centro Studi Auxologici, 2004:175–220.
6 Lee PA, Chernausek SD, Hokken-Koelega ACS, etal. International small for gestational age advisoryboard consensus development conferencestatement: management of short children born smallfor gestational age, April 24–October 1, 2001.Pediatrics 2003;111:1253–61.
7 Nyberg DA, Abuhamad A, Ville Y. Ultrasoundassessment of abnormal fetal growth. SeminPerinatol 2004;28:3–22.
8 Bamberg C, Kalache KD. Prenatal diagnosis of fetalgrowth restriction. Semin Fetal Neonatal Med2004;9:387–94.
9 Deter RL. Individualized growth assessment:evaluation of growth using each fetus as its owncontrol. Semin Perinatol 2004;28:23–32.
10 Ego A, Subtil D, Grange G, et al. Customizedversus population-based birth weight standards foridentifying growth restricted infants: a Frenchmulticenter study. Am J Obstet Gynecol2006;194:1042–9.
11 Cameron N. Measuring techniques andinstruments. In: Nicoletti I, Benso L, Gilli G, eds.Physiological and pathological auxology. Firenze:Edizioni Centro Studi Auxologici, 2004:117–59.
12 Kramer MS, Platt RW, Wen SW, et al. A new andimproved population-based Canadian reference forbirth weight for gestational age. Pediatrics2001;108:E35.
13 Sherry B, Mei Z, Grummer-Strawn L, et al.Evaluation of and recommendations for growthreferences for very low birth weight (,or= 1500 grams) infants in the United States.Pediatrics 2003;11:750–8.
14 Tentoni S, Astolfi P, De Pasquale A, et al.Birthweight by gestational age in preterm babiesaccording to a Gaussian mixture model. BJOG2004;111:31–7.
15 Ogden CL, Kuczmarski RJ, Flegal KM, et al. Centersfor Disease Control and Prevention. 2000 Growthcharts for the United States: improvements to the1977 National Center For Health Statistic version,Pediatrics 2002;109:45–60.
16 Grummer-Strawn LM, Ogden CL, Mei Z, et al.Scientific and pratical issues in the development ofthe US Childhood Growth Reference. In: Martorell R,Haschke F, eds. Nutrition and growth. Philadelphia:Lippincott, 2001:21–36.
17 Bertino E, Murru P, Bagna R, et al. Standardantropometrici neonatali dell’Italia Nord-Occidentale. Riv Ital Ped 1999;25:899–906.
18 Gagliardi L, Macagno F, Pedrotti D, et al. Standardantropometrici neonatali prodotti dalla task-forcedella Societa Italiana di Neonatologia e basati suuna popolazione italiana Nord-Orientale. Riv ItalPed 1999;25:159–69.
19 Bossi A, Milani S. Italian standards for crown-heellength and head circumference at birth. Ann HumBiol 1987;14:321–35.
20 Bernstein IM. The assessment of newborn size.Pediatrics 2003;111:1430–1.
21 World Health Organization. WHO child growthstandards—methods and development. Geneva:WHO, 2006.
22 WHO Multicentre Growth Reference Study Group.Assessment of differences in linear growth amongpopulations in the WHO Multicentre GrowthReference Study. Acta Paediatr 2006;(Suppl450):56–65.
23 Ulijaszek S. Ethnic differences in patterns of humangrowth in stature. In: Martorell R, Haschke F, eds.Nutrition and growth. Philadelphia: Lippincott,2001:1–20.
24 Karlberg J, Cheung YB, Luo ZC. An update on theupdate of growth charts. Acta Paediatr1999;88:797–802.
25 Thomas P, Peabody J, Turnier V, et al. A new lookat intrauterine growth and the impact of race,altitude, and gender. Pediatrics 2000;106:E21.
26 Bertino E, Bagna R, Licata D, et al. Standardantropometrici del neonato da parto bigemino. RivItal Ped 1997;23:98–105.
27 Skiaerven R, Gjessing HK, Bakketeig LS. Newstandards for birth weight by gestational age usingfamily data. Am J Obstet Gynecol2000;183:689–96.
28 McCowan LM, Harding JE, Stewart AW.Customized birthweight centiles predict SGApregnancies with perinatal morbidity. BJOG2005;112:1026–33.
29 Gardosi J. New definition of small for gestationalage based on fetal growth potential. Horm Res2006;65(suppl 3):15–18.
30 Oken E, Kleinman KP, Rich-Edwards J, et al. Anearly continuous measure of birth weight forgestational age using a United States nationalreference. BMC Pediatr 2003;3:6.
31 Gelbaya TA, Nardo LG. Customised fetal growthchart: a systematic review. J Obstet Gynaecol2005;25:445–50.
32 Lemons JA, Bauer CR, Oh W, et al. Very lowbirthweight outcomes of the National Institute ofChild Health and Human Development NeonatalResearch Network January 1995 throughDecember 1996. NICHD Neonatal ResearchNetwork. Pediatrics 2001;107:E1.
33 Rapaport R. Growth and growth hormone inchildren born small for gestational age. GrowthHorm IGF Res 2004;14:S3–6.
34 Monset-Couchard M, De Bethmann O, Relier JP.Long term outcome of small versus appropriatesize for gestational age co-twins/triplets.Arch Dis Child Fetal Neonatal Ed2004;89:F310–14.
35 Levy-Marchal C, Jaquet D, Czernichow P. Long-term metabolic consequences of being born smallfor gestational age. Semin Neonatol2004;9:67–74.
36 Hauspie R, Vercauteren M. Secular trend. In:Nicoletti I, Benso L, Gilli G, eds. Physiological andpathological auxology. Firenze: Edizioni CentroStudi Auxologici, 2004:175–220.
37 Wen SW, Kramer MS, Platt R, et al. Secular trendsof fetal growth in Canada, 1981 to 1997. PaediatrPerinat Epidemiol 2003;17:347–54.
38 Healy MJR, Rasbash J, Yang M. Distribution-freeestimation of age related centiles. Ann Hum Biol1988;15:17–22.
39 Cole TJ. Fitting smoothed centile curves to referencedata. J R Stat Soc 1988;151:385–418.
F10 LEADING ARTICLES
www.archdischild.com
A16
Resuscitation (2006) 71, 319—321
CLINICAL PAPER
Accuracy of clinical assessment of infant heartrate in the delivery room!
C. Omar F. Kamlina,∗, Colm P.F. O’Donnell a,b, Neil J Everesta,Peter G Davisa,b, Colin J Morleya,c
a Division of Newborn Services, Royal Women’s Hospital, Melbourne, Australiab University of Melbourne, Australiac Murdoch Research Institute, Australia
Received 7 February 2006; received in revised form 18 April 2006; accepted 27 April 2006
KEYWORDSInfant;Newborn;Resuscitation;Heart rate
Summary Heart rate (HR) dictates intervention during neonatal resuscitation.Guidelines recommend that HR be assessed by auscultation or palpation. We com-pared HRs determined clinically with electrocardiography (ECG) in healthy newbornsin the delivery room. Clinical assessment by 23 observers randomly allocated toassess HR by one of two methods in 26 infants, was found to be inaccurate andunderestimate ECG HR. The mean difference between HR assessed by auscultationand palpation ECG and HR using methodology recommended by the Neonatal Resus-citation Programme was 14 and 22 beats per minute respectively.© 2006 Elsevier Ireland Ltd. All rights reserved.
Introduction
International consensus statements recommendthat the need for, and response to resuscitationin newborn infants be determined clinically, withthe heart rate being the most important sign.1,2
Counting the heart rate (HR) by auscultation ofthe praecordium with a stethoscope or palpationof umbilical cord pulsations are the recommended
! A Spanish translated version of the summary of thisarticle appears as Appendix in the online version atdoi:10.1016/j.resuscitation.2006.04.015.
∗ Corresponding author at: Research Fellow Neonatal Paedi-atrics, Royal Women’s Hospital Melbourne, 132 Grattan Street,Carlton,Victoria 3053, Australia. Tel.: +61 3 9344 2000 pager2517; fax: +61 3 9344 2185.
E-mail address: [email protected] (C.O.F. Kamlin).
techniques.1,3 Auscultation of the HR has beenshown to be inaccurate in a model,4 while palpa-tion has been shown to be inferior to auscultationin newborns in the delivery room.5 Neither methodhas been compared to electrocardiography (ECG)in the delivery room. We wished to determine theaccuracy of clinical assessment of HR by comparingit to ECG in the delivery room (DR).
Methods
Approval for this study was obtained from theHuman Research and Ethics Committees of theRoyal Women’s Hospital. We identified impendinglow-risk term births and approached parents forconsent prior to delivery. Vigorous infants who were
0300-9572/$ — see front matter © 2006 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.resuscitation.2006.04.015
A17
320 C.O.F. Kamlin et al.
not resuscitated were enrolled. The chest and rightthigh were cleaned using Skin-Prep® wipes (Smith& Nephew, FL USA) and three ECG leads (Kitty-cat, Tyco Healthcare, Mansfield, MA, USA) applied.With the infant on the Resuscitaire, two leads wereapplied to the chest and one to the lateral aspectof the right thigh. The leads were connected toa ProPaq 106EL (Protocol Systems Inc., Oregon,USA) monitor. For each infant, two staff membersattending the birth were invited to participate inthe study once the infant had been dried and sta-bilised. The two care providers were randomly allo-cated (by toss of a coin) to assess HR by eitherpalpation (HRpalp) or auscultation (HRausc). Partici-pants determined the HR in beats per minute (bpm)by multiplying the number of beats heard or felt in6 s by 10 s, as recommended in the Neonatal Resus-citation Programme (NRP).3 Participants simultane-ously determined the HR clinically by each method.They were masked to the ECG HR. Each observerrecorded the HR on a card, which they gave tothe investigator. The HRs assigned by each observerwere compared to the HR measured by ECG. Datawere analysed using SPSS for Windows (Chicago IL,USA) version 11.5 using paired t tests to comparemean HR.
Results
Twenty-three observers (2 consultant neonatolo-gists, 3 neonatal fellows, 7 pediatric registrars, 11neonatal nurses) assessed 26 infants (mean (S.D.)gestational age and birth weight 38 (2) weeks and3191 (681) g, respectively). All infants were activeand no infant received respiratory support.
Figure 1 An error bar chart of the mean difference (95%confidence intervals) between ECG HR and (i) HRausc and(ii) HRpalp.
Comparison of clinical assessment of HR by aus-cultation with ECG recordings were performed in allinfants. However, cord pulsations were impalpableat the time of assessment in 5 (19%) infants. Themean (S.D.) time from birth to application of ECGleads was 1.9 (0.7) min and the mean ECG HR at thetime of clinical evaluation was 168 (21) bpm.
The mean (S.D.) HR from the ECG was 167 (19)bpm compared with HRausc (n = 26) of 154 (22) bpm[p = 0.003]. The mean (S.D.) HR from the ECG was168 (22) bpm compared with HRpalp (n = 21) of 147(19) bpm (p < 0.001). Clinical assessment underes-timated the ECG HR with a mean difference (S.D.)between HRausc, HRpalp and ECG HR of −14 (21) and−21 (21) bpm, respectively (Figure 1).
Discussion
Immediately after birth the new born infant’s HRguides the need for resuscitation.1,2 The best wayto monitor HR at this time has not been determined.This study compared different ways of counting HRin the delivery room in newly born infants whodid not require resuscitation. Vigorous new borninfants were studied because this was the nearestwe could get to investigating the clinical assess-ment of HR accurately immediately after birth.Although it would have been optimal to study HR inthose infants receiving resuscitation it was not pos-sible to do three simultaneous careful assessmentsof HR during resuscitation.
This study has demonstrated inaccuracy in thecounting the HR of newborn infants in the deliv-ery room. Our study did not include infants withHRs <100 bpm, thus we do not know whether clin-ical assessment is similarly inaccurate at slowerHRs. Using a modified infant resuscitation manikin,Theophilopoulos instructed clinicians to deter-mine the heart rate generated by an electronicmetronome using the NRP six second method andtheir own preferred method (duration of listeningvaried from 5 to 30 s).4 Both methods were inaccu-rate at all settings of HR (40, 70, 90 and 120 bpm)but more pronounced when the metronome was setat a HR < 100 bpm, a heart rate at which guidelinessuggest interventions for the infant. Our findings ina clinical setting have shown the tendency for deliv-ery room care providers who are counting time andeither heart beat or pulsations simultaneously, tounderestimate ECG HR across the range of abso-lute ECG HR. If the inaccuracies seen in our andthe manikin study4 were to occur in non-vigorousinfants in the delivery room, it is possible that ther-apies will be either withheld or administered inap-propriately.
A18
Delivery room assessment of infant heart rate 321
Though ECG provides a continuous measure ofHR, obtaining a reading from a wet newborn imme-diately after birth is slow. This study was designedto compare clinical assessment against ECG HR, thegold standard in an intensive care setting. Whilst ithas highlighted the differences between ausculta-tion and ECG HR and the limitations of palpation,costs, access to an ECG on a global scale and timefor data acquisition preclude the use of the ECGto monitor an infant’s HR routinely in the deliveryroom. In contrast, pulse oximetry measures bothoxygen saturation and HR and may be superior tothe ECG in this setting because the readings areeasier to obtain6 but this comparison has not beenassessed. The accuracy of HR assessment by oxime-try immediately after birth needs to be determined.
Clinical trials to determine whether differenttechniques for assessing heart rate, immediatelyafter birth, improve outcomes for resuscitatednewborn infants are the gold standard of scientificassessment but they are difficult to organise andexpensive to run. In the meantime clinicians needto be aware that the clinical assessment of heartrate in the delivery room may be inaccurate.
Conflict of interest statement
None.
Acknowledgements
COFK and CPFO’D both are in receipt of the RoyalWomen’s Hospital Postgraduate Degree Scholarship.PGD is supported by a National Health and MedicalResearch Council Fellowship.
References
1. Niermeyer S, Kattwinkel J, Van Reempts P, et al. Inter-national guidelines for neonatal resuscitation: an excerptfrom the guidelines 2000 for cardiopulmonary resuscita-tion and emergency cardiovascular care: International con-sensus on science. Contributors and reviewers for theneonatal resuscitation guidelines. Pediatrics 2000;106(3):E29.
2. International Liaison Committee on Resuscitation. Part 7:Neonatal Resuscitation. Resuscitation 2005;67:293—303.
3. Kattwinkel J. Neonatal resuscitation textbook. 4th ed. Amer-ican Academy of Pediatrics; 2000.
4. Theophilopoulos DT, Burchfield DJ. Accuracy of differentmethods for heart rate determination during simulatedneonatal resuscitations. J Perinatol 1998;18(1):65—7.
5. Owen CJ, Wyllie JP. Determination of heart rate in the babyat birth. Resuscitation 2004;60(2):213—7.
6. O’Donnell CP, Kamlin CO, Davis PG, Morley CJ. Feasibility ofand delay in obtaining pulse oximetry during neonatal resus-citation. J Pediatr 2005;147(5):698—9.
A19
ORIGINAL ARTICLE
Accuracy of pulse oximetry in assessing heart rate of infants inthe neonatal intensive care unitJasbir KSB Singh,1 C Omar F Kamlin,1 Colin J Morley,1,2,3 Colm PF O’Donnell,1 Susan M Donath3 andPeter G Davis1,2,3
1Division of Paediatrics, Royal Women’s Hospital, 2Department of Obstetrics and Gynaecology, University of Melbourne, and 3Murdoch Children’s ResearchInstitute Melbourne, Melbourne, Victoria, Australia
Aim: To determine the accuracy of pulse oximetry measurement of heart rate in the neonatal intensive care unit.Methods: Stable preterm infants were monitored with a pulse oximeter and an ECG. The displays of both monitors were captured on video.Heart rate data from both monitors, including measures of signal quality, were extracted and analysed using Bland Altman plots.Results: In 30 infants the mean (SD) difference between heart rate measured by pulse oximetry and electrocardiography was -0.4 (12) beatsper minute. Accuracy was maintained when the signal quality or perfusion was low.Conclusion: Pulse oximetry may provide a useful measurement of heart rate in the neonatal intensive care unit. Studies of this technique inthe delivery room are indicated.
Key words: infant; newborn; pulse oximetry.
Measurement of heart rate (HR) is an important component ofthe assessment of neonates both in the neonatal intensive careunit (NICU) and in the delivery room (DR). Electrocardiography(ECG) is the accepted ‘gold standard’ method of continuouslymonitoring HR and has been used in the NICU for decades.However, ECG chest leads may cause breakdown of the fragileskin of extremely preterm infants. Pulse oximetry (PO) has alsobeen used in the NICU for many years and provides a continu-ous, non-invasive measurement of oxygen saturation and HR.
In the DR, HR is traditionally assessed clinically, either bypalpation of the umbilical cord or by auscultation of the heart.1
Both methods are inaccurate, systematically underestimatingHR by an average of 14–21 beats per min (bpm).2 As all newlyborn infants are wet, application of ECG monitoring in the DRis technically difficult and time-consuming. With the availabilityof new generation monitors that cope better with motion arte-fact, it has been suggested that PO may have a role in the DR toassist both optimising oxygenation and measuring HR. POmonitoring in the DR may be more difficult than in the NICU
because of motion of the infant and poor peripheral perfusion inthe first minutes of life. New generation pulse oximeters displayboth a low signal identification and quality indicator (signal IQ),when data may be erroneous because of excessive motion orlow perfusion and a perfusion index (PI) which provides arelative assessment of pulse strength at the monitoring site.
The accuracy of PO measurement of HR in the NICU and inthe DR is not well studied. If inaccurate in the NICU, it isunlikely that the technique would be useful in the DR. There-fore, in an observational study we sought to determine theaccuracy of PO in the NICU and how this varied when themonitor displayed low signal IQ or PI.
Methods
Stable preterm infants who were monitored with ECG in theNICU were eligible for inclusion. Basic demographic data werecollected on each baby. A Masimo Radical (Masimo Corpora-tion, Irvine, CA, USA) PO monitor was placed next to the ECGmonitor. The PO sensor was placed on the right hand, securedwith Coban wrap (3M Healthcare, St. Paul, MN, USA) and thenconnected to the Masimo Radical PO cable as previouslydescribed.3 The PO was set to acquire data with maximal sensi-tivity and average it over 2-s intervals. A Panasonic NVGS200digital video camera (Matushita Corporation, Japan) mountedon a Manfrotto Magic Arm (Manfrotto, Italy) was used to recordthe displays of both monitors.
The videos obtained were reviewed and data were extractedevery 2 s on HR, PI and signal IQ. Data extraction began fromthe time when both PO and ECG data were available and con-tinued for 180 s.
Key Points
1 Pulse oximetry accurately measures heart rate in the NICU.2 Accuracy is maintained even when signal quality is low.3 Further studies are required to determine the role of pulse
oximetry in the delivery room.
Correspondence: Associate Professor Peter G Davis, Department ofObstetrics and Gynaecology, The Royal Women’s Hospital, 132 GrattanStreet, Carlton, Vic. 3053, Australia. Fax: +61 3 9347 1761; email:[email protected]
Accepted for publication 31 August 2007.
doi:10.1111/j.1440-1754.2007.01250.x
Journal of Paediatrics and Child Health 44 (2008) 273–275© 2007 The AuthorsJournal compilation © 2007 Paediatrics and Child Health Division (Royal Australasian College of Physicians)
273
A20
A Bland Altman plot was constructed to assess the level ofagreement between the ECG HR and the PO HR. This methodplots the difference between these two values against theiraverage.4 Two standard deviations around the mean differencerepresented the limit of agreement.4 Similar plots were con-structed for the subset of Masimo HR values with (i) low PI and(ii) low signal quality. The Human Research and Ethics Com-mittees of The Royal Women’s Hospital, Melbourne, Australiaapproved the study.
Results
A convenience sample of 30 stable infants in the NICU wasstudied. The mean (SD) gestational age was 29 (4) weeks andbirthweight 1451 (905) g. A total of 2730 pairs of PO and ECGHRs were analysed. The Bland Altman plot showed the meandifference between PO and ECG HR was -0.4 bpm and 2 SD was12 bpm (Fig. 1). In the subset of PO values with the lowest 5%PI (<0.47) the mean difference was identical to the overall resultand 2 SD was 14.4 bpm (Fig. 2). In the subset of readings wherethe PO displayed low signal quality the mean difference was0.9 bpm and 2 SD was 28.4 bpm (Fig. 3).
Discussion
Reports of the usefulness of PO to monitor infants in the inten-sive care unit and under anaesthesia appeared in the 1980s.5,6
Case series illustrating its application for monitoring infants inthe DR appeared shortly afterwards.7,8 These early observationalstudies noted the limitations of the technique in situationswhere the infant was moving or had low perfusion. The devel-opment of signal extraction technology enabled continuous datato be obtained for a greater proportion of time than older gen-eration oximetry.9
Evaluation of PO in both the NICU and DR has focusedpredominantly on the measurement of oxygen saturation. The
ability to measure HR has largely been overlooked. However,Barrington et al. tested a variety of pulse oximeters available inthe late 1980s, comparing the HR with ECG.10 They found thatPO HR deviated from ECG HR by more than 10 bpm for 11.9–25% of the time. Hay et al. assessed the accuracy of a newgeneration signal extraction technology monitor in detectingbradycardic events in the NICU patients and found it to besuperior to other monitors.11 Our study confirms the accuracy ofPO measurement of HR in the NICU.
We suggest two novel settings in which measurement of HRby PO may be useful. First, in newborn premature infantswith fragile skin who are unable to tolerate conventional ECG
200180160140120100
Average heart rate [ ( Masimo + ECG ) / 2 ] bpm
60
40
20
0
20
40Hea
rt r
ate
diffe
renc
e (
Mas
imo
Mean difference ( SD ) =
+2 SD
2 SD
EC
G )
bpm
0.4 ( 6.0 ) bpm
Fig. 1 Bland Altman plot of heart rate difference versus average heartrate (all values). bpm, beats per minute; ECG, electrocardiography.
190180170160150140130120
Average heart rate [ ( Masimo + ECG ) / 2 ] bpm
40
30
20
10
0
10
20Hea
rt r
ate
diffe
renc
e (
Mas
imo
Mean difference ( SD ) =
+2 SD
2 SD
EC
G )
bpm
0.4 ( 7.2 ) bpm
Fig. 2 Bland Altman plot for Masimo pulse oximetry heart rate valueswith lowest 5% perfusion (perfusion index < 0.47). bpm, beats per minute;ECG, electrocardiography.
190180170160150140130
Average heart rate [ ( Masimo + ECG ) / 2 ] bpm
60
40
20
0
20
40Hea
rt r
ate
diffe
renc
e (
Mas
imo
Mean difference ( SD ) = 0.9 ( 14.2 ) bpm
+2 SD
2 SD
EC
G )
bpm
Fig. 3 Bland Altman plot for Masimo heart rate values with lowsignal identification and quality indicator. bpm, beats per minute; ECG,electrocardiography.
Pulse oximetry to measure heart rate JKSB Singh et al.
Journal of Paediatrics and Child Health 44 (2008) 273–275© 2007 The Authors
Journal compilation © 2007 Paediatrics and Child Health Division (Royal Australasian College of Physicians)
274
A21
monitoring, PO may be used to measure HR. Second, an alter-native to the intermittent, inaccurate measurement of HR bypalpation or auscultation in the DR is required. Assessment ofthe condition of newly born infants and their response to treat-ment has remained largely unchanged for half a century. POprovides continuous, accurate assessment of HR of stable infantsin the NICU, even when PI and signal quality are low. Extensionof its use to the DR offers a potential solution to a long-standingproblem. Further studies are required before PO is recom-mended in either setting.
References
1 The International Liaison Committee on Resuscitation (ILCOR)consensus on science with treatment recommendations for pediatricand neonatal patients: neonatal resuscitation. Pediatrics 2006; 117:e978–88.
2 Kamlin CO, O’Donnell CP, Everest NJ, Davis PG, Morley CJ. Accuracyof clinical assessment of infant heart rate in the delivery room.Resuscitation 2006; 71: 319–21.
3 O’Donnell CP, Kamlin CO, Davis PG, Morley CJ. Obtaining pulseoximetry data in neonates: a randomised crossover study of sensor
application techniques. Arch. Dis. Child. Fetal Neonatal Ed. 2005; 90:F84–5.
4 Bland JM, Altman DG. Statistical methods for assessing agreementbetween two methods of clinical measurement. Lancet 1986; 1:307–10.
5 Brooks TD, Gravenstein N. Pulse oximetry for early detection ofhypoxemia in anesthetized infants. J. Clin. Monit. 1985; 1: 135–7.
6 Wasunna A, Whitelaw AG. Pulse oximetry in preterm infants.Arch. Dis. Child. 1987; 62: 957–8.
7 Maxwell LG, Harris AP, Sendak MJ, Donham RT. Monitoring theresuscitation of preterm infants in the delivery room using pulseoximetry. Clin. Pediatr. (Phila) 1987; 26: 18–20.
8 Dimich I, Singh PP, Adell A, Hendler M, Sonnenklar N, Jhaveri M.Evaluation of oxygen saturation monitoring by pulse oximetry inneonates in the delivery system. Can. J. Anaesth. 1991; 38:985–8.
9 Kopotic RJ, Lindner W. Assessing high-risk infants in the delivery roomwith pulse oximetry. Anesth. Analg. 2002; 94: S31–6.
10 Barrington KJ, Finer NN, Ryan CA. Evaluation of pulse oximetry as acontinuous monitoring technique in the neonatal intensive care unit.Crit. Care Med. 1988; 16: 1147–53.
11 Hay WW Jr, Rodden DJ, Collins SM, Melara DL, Hale KA, Fashaw LM.Reliability of conventional and new pulse oximetry in neonatalpatients. J. Perinatol. 2002; 22: 360–6.
JKSB Singh et al. Pulse oximetry to measure heart rate
Journal of Paediatrics and Child Health 44 (2008) 273–275© 2007 The AuthorsJournal compilation © 2007 Paediatrics and Child Health Division (Royal Australasian College of Physicians)
275
A22
ORIGINALARTICLES
Accuracy of Pulse Oximetry Measurement of Heart Rate of NewbornInfants in the Delivery Room
C. OMAR F. KAMLIN, MBBS, MRCPCH, JENNIFER A. DAWSON, BSC, COLM P. F. O’DONNELL, MRCPCH, PHD,COLIN J. MORLEY, MD, FRACP, SUSAN M. DONATH, PHD, JASBIR SEKHON, AND PETER G. DAVIS, MD, FRACP
Objective To determine the accuracy of heart rate obtained by pulse oximetry (HRPO) relative to HR obtained by 3-leadelectrocardiography (HRECG) in newborn infants in the delivery room.Study design Immediately after birth, a preductal PO sensor and ECG leads were applied. PO and ECG monitor displayswere recorded by a video camera. Two investigators reviewed the videos. Every two seconds, 1 of the investigators recordedHRPO and indicators of signal quality from the oximeter while masked to ECG, whereas the other recorded HRECG and ECGsignal quality while masked to PO. HRPO and HRECG measurements were compared using Bland-Altman analysis.Results We attended 92 deliveries; 37 infants were excluded due to equipment malfunction. The 55 infants studied had amean (!standard deviation [SD]) gestational age of 35 (!3.7) weeks, and birth weight 2399 (!869) g. In total, we analyzed5877 data pairs. The mean difference (!2 SD) between HRECG and HRPO was "2 (!26) beats per minute (bpm) overall and"0.5 (!16) bpm in those infants who received positive-pressure ventilation and/or cardiac massage. The sensitivity andspecificity of PO for detecting HRECG <100 bpm was 89% and 99%, respectively.Conclusion PO provided an accurate display of newborn infants’ HR in the delivery room, including those infants receivingadvanced resuscitation. (J Pediatr 2008;152:756-60)
An infant’s heart rate (HR) is used to assess the need for and response toresuscitation.1 HR is determined by auscultating the precordium or palpating theumbilical cord. Intervention is recommended if the HR is !100 beats per minute
(bpm).2,3 Clinical assessment of HR in the delivery room is intermittent and ofteninaccurate.4,5 Pulse oximetry (PO), increasingly used in the delivery room,6 allowstitration of oxygen concentrations delivered to infants receiving respiratory support.7 Wehave demonstrated that HR obtained by PO (HRPO) in the neonatal intensive care settingis accurate compared with HR obtained from 3-lead electrocardiography (HRECG).8
ECG monitoring in the delivery room is difficult due to the infant’s wet skin, istime-consuming,4 and may damage the skin of an extremely preterm infant. ConventionalPO is unreliable in the delivery room, where excessive motion and low perfusion can leadto underestimation of true HR.9 Recently developed second-generation pulse oximetershave advanced processing algorithms to adjust for patient motion and poor perfusionstates, thereby improving accuracy and precision.10,11 PO data are easily obtained fromnewborn infants in the delivery room,12 but the accuracy of the HRPO is unknown.
In the present study, we compared HRPO determined by second-generation pulseoximetry with HRECG in newborn infants in the delivery room. Specifically, we sought toevaluate the precision and accuracy of PO across a wide range of HRECG readings and thesensitivity and specificity of PO in detecting HRECG !100 bpm.
METHODSTwo investigators attended the deliveries of a convenience sample of newborn
infants between July 2005 and August 2006. Use of the pulse oximeter in the delivery
bpm Beats per minuteECG ElectrocardiographyHR Heart rate
PO Pulse oximetrySD Standard deviation
See editorial, p 747 andrelated article, p 761
From the Neonatal Services, The RoyalWomen’s Hospital, Melbourne, Australia(C.K., J.D., C.M., P.D.); Neonatal Unit, Na-tional Maternity Hospital, Dublin, Ireland(C.O.); Department of Obstetrics, Univer-sity of Melbourne, Melbourne, Australia(C.M., S.D., J.S., P.D.); and the MurdochChildren’s Research Institute, Melbourne,Australia (C.M., S.D.).Supported in part by a RWH Postgraduatedegree scholarship (C.O., C.K., J.D.) anNHMRC Practitioner Fellowship (P.D.),and an Australian National Health andMedical Research Council Program grant(384100).Submitted for publication Jul 16, 2007; lastrevision received Oct 25, 2007; acceptedJan 3, 2008.Reprint requests: C. Omar F. Kamlin, Neo-natal Paediatrician, Division of NewbornServices, The Royal Women’s Hospital, 132Grattan St, Carlton, Victoria, Australia,3053. E-mail: [email protected]/$ - see front matterCopyright © 2008 Mosby Inc. All rightsreserved.10.1016/j.jpeds.2008.01.002
756
A23
room has become a standard of care at our institution, andapproval for the study as an audit of practice was obtainedfrom the Royal Women’s Hospital, Melbourne Research, andHuman Research Ethics Committees. Verbal consent forvideo recording of the 2 monitors in the delivery room wasobtained from the infant’s parents before delivery.
Immediately after birth, the infant was placed on aresuscitation trolley, where a PO sensor (L-NOP Neo;Masimo Corp, Irvine, CA) was applied to the infant’s righthand/wrist and then connected to a Masimo Radical signal-extraction pulse oximeter.12,13 The oximeter was set to ac-quire data with maximum sensitivity and a 2-second averag-ing interval.
Simultaneously, another study investigator applied 3ECG chest leads (Puppydog; Kendall, Mansfield, MA), con-nected to an Escort II ECG monitor (MDE, Arleta, CA).The HR was based on a measure of the interval between Rwaves, averaged over 6 complexes. The oximeter and ECGmonitor were placed side by side, and both screens wererecorded by a video camera. PO and ECG measurementswere available to the clinical team responsible for the infant.
Two investigators independently reviewed the videos.The videos were paused every 2 seconds to record data. Thestarting point for data collection was 10 seconds from thetime at which both the HRPO and HRECG data were dis-played. One investigator recorded HRPO, perfusion index(PI), and indicators of signal quality from the oximeter whilemasked to the ECG. The signal was deemed to be of “poor”quality if signal bars were absent at the bottom of the screenand/or the message “low-signal IQ” was displayed. The pres-ence of bars and absence of message indicated a “good”-quality signal irrespective of the shape of the plethysmo-graph.14 The second investigator recorded the HRECG andsignal quality while masked to the PO. The ECG trace wasdeemed to be of good quality if the QRS complexes were wellformed and the rhythm strip was consistent with the HRECGdisplayed on the monitor. Data were recorded for a minimumof 3 minutes (90 data points). For an infant receiving ad-vanced interventions (positive-pressure ventilation by facemask or greater), data collection continued until the infantwas stabilized.
The relationship between HRPO and HRECG was eval-uated using Bland-Altman bias analysis; the difference be-tween the measurements was plotted against their average.Two standard deviations around the mean difference repre-
sented the upper and lower limits of agreement.15,16 Thesensitivity, specificity, positive predictive values, and negativepredictive values of PO for detecting HRECG !100 bpm werecalculated. The data were analyzed using intercooled Stataversion 9.2 (StataCorp, College Station, TX).
RESULTSWe attended 92 deliveries and failed to make recordings
at 37 of them, due to 20 ECG, 12 PO, and 5 video cameramalfunctions. Consequently, we recorded and analyzed datafrom 55 infants who had a mean ("standard deviation [SD])birth weight of 2399 g ("876 g) and a mean gestational ageof 35 weeks ("3.7 weeks). The resuscitation administered tothese infants is detailed in the Table. The time (median[interquartile range]) taken to acquire HRPO was 68 seconds(60 to 118 seconds); that to acquire HRECG was 80 seconds(64 to 104 seconds); and that to the start of data collectionwas 118 seconds (90 to 155 seconds).
A total of 6475 HRECG and 6448 HRPO data pointswere entered into the spreadsheet; the difference was due totransient loss of HRPO and/or HRECG. All of the HRPO datapoints were used to determine levels of agreement where theHRECG was deemed to be of good quality (n # 5877). TheHRPO and HRECG data for all infants were compared in aBland-Altman plot (Figure 1A). The mean difference (HRPO- HRECG) was $2 bpm, and the 95% limit of agreement ("2SD) was "26 bpm. Figure 1B shows data from 2 active terminfants in whom HRPO significantly underestimated HRECG.When HRECG was compared with good-quality HRPO (ie,presence of signal bars and absence of a “low-signal IQ”message; n# 5143), the mean difference ("2 SD) was $1bpm ("20 bpm).
Figure 2 displays the level of agreement between HRPOand HRECG for those infants receiving advanced resuscita-tion. The mean difference (2 SD) for the 5 infants who wereintubated (Figure 2A) and the 2 infants who received externalcardiac massage (Figure 2B) was 0.2 ("16) and 0.8 ("8)bpm, respectively.
Good-quality signals from both devices (n # 5143) arecompared in Figure 3. A strong linear correlation (r2 # 0.8)can be seen. Dividing this graph into 4 quadrants, above andbelow a HR of 100 bpm revealed a sensitivity of 89%, spec-ificity of 99%, positive predictive value of 83%, and negativepredictive value of 99% of PO to detect a HRECG !100 bpm.A HRECG !100 bpm was detected by PO 89% of the time.
Table. Delivery room interventions received by infants (not mutually exclusive)
Delivery room intervention
Gestational age, weeks
Total (%)28 to 31 (n # 12) 32 to 35 (n # 15) >36 (n # 28)
None 4 8 24 36 (65)Supplemental oxygen 8 8 3 19 (35)Continuous positive airway pressure by mask 6 6 1 13 (24)Positive-pressure ventilation by mask 6 6 0 12 (22)Positive-pressure ventilation by endotracheal tube 3 2 0 5 (9)Chest compressions 1 1 0 2 (4)
Accuracy of Pulse Oximetry Measurement of Heart Rate of Newborn Infants in the Delivery Room 757
A24
There were 42 data points from 10 infants in which HRPOwas !100 bpm when HRECG was %100 bpm. One of these10 infants contributed 16 data points (32 seconds); this infantreceived positive-pressure ventilation. The other 9 infantseach contributed between 1 and 4 data points.
DISCUSSIONIn this study, we evaluated the accuracy and precision of
HRPO in the delivery room using a new-generation pulseoximeter. Our results suggest that in the delivery room,HRPO is accurate and on average within 2 bpm of HRECG.Overall, the 95% limit of agreement was wider than what weexpected, at "26 bpm. However, in the infants who receivedthe most intervention in our cohort, the level of agreementwas "16 bpm. This compares with the pulse oximeter man-ufacturer’s level of agreement of "10 bpm when tested onadult subjects with simulated motion.14
A weakness of our study is the high number of excludedinfants (40%). The most common technical problem wasfailure to acquire HRECG. The video camera failed during 5deliveries, due to lack of power or videotape. The pulseoximeter battery failed during 5 deliveries, and in 7 vigorousinfants there was poor signal acquisition due to motion arte-fact. PO failed to acquire data in this setting in 12 of 92infants (13%). Consequently, we advise caution in the appli-cation of this new technology; it may fail or, equally impor-tantly, divert attention away from management of the infantin the stressful environment of the delivery room.
ECG and PO measure HR differently. There are dif-ferences in averaging intervals between the 2 devices, and thusit is possible that we have underestimated the true agreementbetween PO and ECG. With rapidly changing heart ratesduring neonatal transition, we observed that HRPO may lagbehind HRECG by a few seconds. Although clinically unim-
Figure 1. Bland-Altman plot showing the level of agreement between HRPO and HRECG (when HRECG signals were deemed to be of good quality) inall infants (A) and in a subgroup of 2 infants in whom HRECG was consistently underestimated (B).
Figure 2. Bland-Altman plot showing the level of agreement between HRPO and HRECG (when signals from both devices were deemed to be of goodquality) based on maximum resuscitation intervention in the 5 infants who were intubated (A) and the 2 infants who received external cardiac massage (B).
758 Kamlin et al The Journal of Pediatrics • June 2008
A25
portant, this lag would introduce differences between the 2monitors in our data set, as demonstrated by the 26 readingsin 5 infants in whom HRPO remained %100 bpm while theHRECG had dropped to !100 bpm (Figure 3). One infant, of31 weeks’ gestation, contributed 12 data points in which theranges of readings from the PO and ECG were 101 to 104bpm and 96 to 99 bpm, respectively. These recordings arewithin the margin of error of the devices and are clinicallyunimportant.
The results on a small number of infants who requiredadvanced resuscitation demonstrate the potential of this tech-nology under these circumstances. Precision and accuracywere slightly improved in this situation (Figure 2). False-positive HRPO readings (ie, where the PO incorrectly iden-tifies an infant as having a low HR, potentially leading toinappropriate intervention) may be a concern. In our sample,42 data points from 10 infants were noted in which HRPOwas !100 bpm and HRECG was %100 bpm (Figure 3). Forall but 1 infant, these readings lasted between 2 and 8 sec-onds—insufficiently long, we believe, to alter management. In1 infant who contributed 16 false-positive data points (32seconds), PO may have led to interventions that were notindicated. However, overtreatment is probably less likelywhen using HRPO than when using auscultation, which sys-tematically underestimates HR by 17 bpm.4
PO may be inaccurate in vigorous infants. In our study,there were 2 infants in whom HRPO underestimated HRECGby 100 to 130 bpm (Figure 1B). This may be due to poorapposition of the light-emitting diode and the photodetectorafter excessive limb movement or reduced perfusion as a resultof fist-clenching. These infants were not excluded from theanalysis because of our a priori commitment to record and useonly PI, “low-signal IQ,” and the presence of signal bars asindicators of signal quality.
The International Committee on Resuscitation recom-mends 2 methods of evaluating HR: auscultation of theprecordium and palpation of the umbilical cord.2 Although
auscultation is superior to palpation, both methods are inter-mittent and inaccurate.4,5 We recognize that not every birth-ing area will have access to PO and that the stethoscope willremain the main instrument for assessing newborn infantsworldwide. Based on our findings, we recommend using POas an adjunct to careful clinical surveillance of infant HR,especially for very preterm infants who may need resuscita-tion. We have shown that PO can identify with high sensi-tivity and specificity those infants who require interventionbased on current recommendations.
The availability and adoption of PO in the deliveryroom is increasing worldwide without extensive previous in-vestigation. Kopotic and Lindner10 examined the feasibilityand utility of the Masimo signal-extraction pulse oximeter inthe delivery room to titrate the fraction of inspired oxygen totarget saturations and assess the need for intervention usingHR. In a feasibility study of early continuous positive airwaypressure for extremely low birth weight infants in the deliveryroom, Finer et al used HRPO to guide intervention in all but1 of 47 infants who were intubated.17 Other studies haveinvestigated the effects of endotracheal intubation on HR andperipheral oxygen saturation in the delivery room;18 none ofthese studies was designed to verify the accuracy of HRPO,however. Our data demonstrate that HRPO is of sufficientaccuracy to be of use to clinicians and researchers alike.
The accuracy and precision of this technology have not yetbeen studied in extremely preterm infants; studies are needed inthis patient group. Future studies also should address the ques-tion of whether PO used as an adjunct to clinical evaluationimproves important patient outcomes in the delivery room.
REFERENCES1. Saugstad OD, Ramji S, Rootwelt T, Vento M. Response to resuscitation of thenewborn: early prognostic variables. Acta Paediatr 2005;94:890-5.2. The International Liaison Committee on Resuscitation (ILCOR) consensus onscience with treatment recommendations for pediatric and neonatal patients: neonatalresuscitation. Pediatrics 2006;117:e978-88.3. Paediatric advanced life support: Australian Resuscitation Council guidelines2006. Emerg Med Australas 2006;18:357-71.4. Kamlin CO, O’Donnell CP, Everest NJ, Davis PG, Morley CJ. Accuracy ofclinical assessment of infant heart rate in the delivery room. Resuscitation 2006;71:319-21.5. Owen CJ, Wyllie JP. Determination of heart rate in the baby at birth. Resusci-tation 2004;60:213-7.6. O’Donnell CP, Davis PG, Morley CJ. Positive pressure ventilation at neonatalresuscitation: review of equipment and international survey of practice. Acta Paediatr2004;93:583-8.7. Kattwinkel J. Textbook of Neonatal Resuscitation. 5th edition. Elk Grove, IL:American Academy of Pediatrics and American Heart Association; 2006.8. Sekhon JSK, Kamlin CO, Morley CJ, O’Donnell CP, Donath S, Davis PG.Accuracy of pulse oximetry in assessing heart rate of infants in the neonatal intensive careunit. J Paediatr Child Health, in press.9. Meier-Stauss P, Bucher HU, Hurlimann R, Konig V, Huch R. Pulse oximetryused for documenting oxygen saturation and right-to-left shunting immediately afterbirth. Eur J Pediatr 1990;149:851-5.10. Kopotic RJ, Lindner W. Assessing high-risk infants in the delivery room withpulse oximetry. Anesth Analg 2002;94(1 Suppl):S31-6.11. Sahni R, Gupta A, Ohira-Kist K, Rosen TS. Motion-resistant pulse oximetry inneonates. Arch Dis Child Fetal Neonatal Ed 2003;88:F505-8.12. O’Donnell CP, Kamlin CO, Davis PG, Morley CJ. Feasibility of and delay inobtaining pulse oximetry during neonatal resuscitation. J Pediatr 2005;147:698-9.13. O’Donnell CP, Kamlin CO, Davis PG, Morley CJ. Obtaining pulse oximetry datain neonates: a randomised crossover study of sensor application techniques. Arch DisChild Fetal Neonatal Ed 2005;90:F84-5.
Figure 3. Scatterplot of paired data points when the signals from bothdevices were deemed to be of good quality. Each circle represents 1 datapair. The data are divided into 4 quadrants. HRECG (x axis) # 100 bpm;HRPO (y axis) # 100 bpm.
Accuracy of Pulse Oximetry Measurement of Heart Rate of Newborn Infants in the Delivery Room 759
A26
14. Operator’s manual: Radical signal-extraction pulse oximeter. Irvine, CA: MasimoCorp; 2004.15. Bland JM, Altman DG. Statistical methods for assessing agreement between twomethods of clinical measurement. Lancet 1986;8476:307-10.16. Bland JM, Altman DG. Measuring agreement in method comparison studies. StatMethods Med Res 1999;8:135-60.
17. Finer NN, Carlo WA, Duara S, Fanaroff AA, Donovan EF, Wright LL, et al.Delivery room continuous positive airway pressure/positive end-expiratory pressure inextremely low birth weight infants: a feasibility trial. Pediatrics 2004;114:651-7.18. O’Donnell CP, Kamlin CO, Davis PG, Morley CJ. Endotracheal intubationattempts during neonatal resuscitation: success rates, duration, and adverse effects.Pediatrics 2006;117:e16-21.
50 Years Ago in The Journal of PediatricsRECENT ADVANCES IN GENETICS IN RELATION TO PEDIATRICS
Fraser FC. J Pediatr 1958;52:734-57
Five years after Watson and Crick published their hypothesis on the structure of DNA,1 Fraser opined about thetrends in genetics, as well as their potential clinical applications. It is astounding how far the field has come in half acentury. We now know that the normal chromosomal complement is 46, not 48, and that DNA is the carrier of geneticinformation. What is currently called the “central dogma” (DNA¡RNA¡ protein) was a mere “modern concept” 50years ago. The Human Genome Project, which completed the sequencing of the human genome in 2003, was not evena twinkle in a geneticist’s eye in 1958.
From conceptual foundations to practical applications, the role of genetics in medicine has bloomed. Now, moleculartesting is the gold standard for diagnosing a multitude of genetic conditions and is the preferred modality for assessingcarrier status. State-coordinated newborn screening programs began in the 1960s with testing for phenylketonuria usinga dried blood spot. Recently, many states have expanded their panels to screen for more than 40 conditions as a resultof the introduction of tandem mass spectrometry into these programs, which has dramatically improved the ability todiagnose and treat inborn errors of metabolism before symptoms develop.
The first genetic counseling graduate program was established a decade after Fraser noted that “the demand for geneticcounseling is growing and there is an increasing need for suitably trained counselors.” There are now more than 2400board-certified genetic counselors, and the need for these professionals continues to transcend multiple specialties,including pediatrics, adults, prenatal, cancer, cardiovascular, hematology, and neurology.
Most fascinating is how Fraser’s words still ring true today: “Exciting things have been happening in genetics in thepast few years, many of them directly or potentially relevant to the practice of medicine.” This discipline has progressedgreatly in the last 50 years, and there is still plenty of excitement to come. Medical Genetics is now a primary specialtywith more than 1000 board-certified clinical geneticists, “personalized medicine” is on the horizon, and genetic testingis being marketed directly to consumers. The questions we now face are not just how to identify genes, but rather whatcan/should we do with this information. It was surely impossible for Fraser to predict that genetics would be so intimatelyinvolved in all aspects of health care, as we now know it today.
Angela Lanie Duker, MS, CGCGenetic Counselor
Division of Genetic and Metabolic DisordersChildren’s Hospital of Michigan
Detroit, Michigan
Gerald L. Feldman, MD, PhD, FACMG, FAAPDivision of Genetic and Metabolic Disorders/Children’s Hospital of Michigan
Professor, Pediatrics, Pathology and Center for Molecular Medicine and GeneticsWayne State University School of Medicine
Detroit, Michigan10.1016/j.jpeds.2007.12.038
REFERENCE1. Watson JD, Crick FH. Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid. Nature 1953;171:737-8.
760 Kamlin et al The Journal of Pediatrics • June 2008
A27
Oxygen saturation and heart rate during deliveryroom resuscitation of infants ,30 weeks’ gestationwith air or 100% oxygenJ A Dawson,1,2 C O F Kamlin,1 C Wong,1 A B te Pas,1 C P F O’Donnell,3 S M Donath,4
P G Davis,1,2 C J Morley1,2,5
1 Neonatal Services, The RoyalWomen’s Hospital, Melbourne,Australia; 2 Department ofObstetrics and Gynaecology,University of Melbourne,Australia; 3 National MaternityHospital, Dublin, Ireland;4 Clinical Epidemiology andBiostatistics Unit, MurdochChildren’s Research Institute,Melbourne, Australia; 5 MurdochChildren’s Research Institute,Melbourne, Australia
Correspondence to:Jennifer Dawson, NeonatalServices, The Royal Women’sHospital, 20 Flemington Road,Parkville, VIC 3052, Australia;[email protected]
Accepted 27 July 2008Published Online First14 August 2008
ABSTRACTBackground: Because of concerns about harmful effectsof 100% oxygen on newborn infants, air has started to beused for resuscitation in the delivery room.Objective: To describe changes in preductal oxygensaturation (SpO2) and heart rate (HR) in the first 10 minafter birth in very preterm infants initially resuscitatedwith 100% oxygen (OX100) or air (OX21).Patients and methods: In July 2006, policy changedfrom using 100% oxygen to air. Observations of SpO2 andHR before and after the change were recorded whenevera member of the research team was available to attendthe birth.Results: There were 20 infants in the OX100 group and106 in the OX21 group. In the OX100 group, SpO2 had risento a median of 84% after 2 min and 94% by 5 min. In theOX21 group, median SpO2 was 31% at 2 min and 54% at5 min. In the OX21 group, 92% received supplementaloxygen at a median of 5 min; the SpO2 rose to a medianof 81% by 6 min. In the first 10 min after birth, 80% and55% of infants in the OX100 and OX21 groups, respectively,had an SpO2 >95%. Increases in HR over the first 10 minwere very similar in the two groups.Conclusions: Most very preterm infants receivedsupplemental oxygen if air was used for the initialresuscitation. In these infants, the use of backup 100%oxygen and titration against SpO2 resulted in a similarcourse to ‘‘normal’’ term and preterm infants. Of theinfants resuscitated with 100% oxygen, 80% had SpO2>95% during the first 10 min. The HR changes in the twogroups were very similar.
For many years, 100% oxygen was recommendedfor delivery room (DR) resuscitation of newborninfants of all gestational ages.1 In recent years,experts have suggested that even a brief exposureto high oxygen concentrations at birth in very-low-birthweight infants is harmful.2 There is alsoaccumulating evidence of oxygen toxicity fromanimal and in vitro studies.3–5 Several studies havefound evidence of oxidative damage in infants aftershort exposure to 100% oxygen in the DR.6–9 Thisevidence has led to a change in national guidelinesfor DR resuscitation, which now advise that 21%oxygen should be considered rather than 100%oxygen during initial resuscitation of all infants.10 11
Very preterm infants appear to be at greatest riskof oxidative damage.7 9 12 Data from infants born at,30 weeks’ gestation are limited.The protocol for resuscitation of newborn infants
at The Royal Women’s Hospital, Melbourne waschanged in line with recommendations of the
Australian Resuscitation Council.10 Before thechange, 100% oxygen was used throughout resusci-tation. After the change, air was used with 100%oxygen as a backup guided by ranges of oxygensaturations previously seen in term and preterminfants not requiring resuscitation.13 14 We aimed todescribe the changes in oxygen saturation (SpO2) andheart rate (HR) of infants ,30 weeks’ gestationresuscitated in the two eras.
PATIENTS AND METHODSIn this prospective observational study, the cohortsincluded infants,30 weeks’ gestation, born at TheRoyal Women’s Hospital, Melbourne between 12January 2006 and 31 December 2007, when amember of the research team was available toattend the birth. This included 6 months before(OX100) and 18 months after (OX21) 17 July 2006,when the change in policy and implementation ofthe new guideline took place.Immediately after birth, an oximetry sensor
(LNOP Neo sensor; Masimo, Irvine, California,USA) was placed on the infant’s right hand andthen connected to the oximeter (Radical7 V5;Masimo), as previously described.15 SpO2, HR andsignal quality were stored by the oximeter every2 s for at least the first 10 min after birth. We used
What is already known on this topic
c Brief exposure to supplemental oxygen in thedelivery room can produce an SpO2 >95%.
c Preductal oximetry measures SpO2 and heartrate within 90 s of birth.
c Using oximetry in the delivery room, clinicianscan adjust the FiO2 to the SpO2.
What this study adds
c When very preterm infants are initiallyresuscitated with air, some will requiresupplemental oxygen.
c If very preterm infants are initially resuscitatedwith 100% oxygen, many will rapidly becomehyperoxic.
c Titrating oxygen administration may reduce thenumber of very preterm infants with SpO2measurements >95%.
Original article
Arch Dis Child Fetal Neonatal Ed 2009;94:F87–F91. doi:10.1136/adc.2008.141341 F87
A28
2 s averaging and maximum sensitivity. A member of theresearch team documented any interventions during resuscita-tion including adjustments made to the fractional inspiredoxygen (FiO2). In the OX100 group, the FiO2 was not able to bechanged because there was no gas blender available until infantswere moved into a transport cot for transfer to the neonatalintensive care unit. Infants in the OX21 group were managedaccording to the 2006 Royal Women’s Hospital DR protocolwith oxygen titrated according to SpO2 measurements (fig 1). Ifinfants reached an SpO2 .90%, the FiO2 was reduced in stages of,10% to target the SpO2 to 80–90%. Active, spontaneouslybreathing infants, of any gestation, were started on continuouspositive airway pressure (CPAP); infants requiring additionalsupport were intubated and ventilated. This did not change overthe time of this study. If free flow oxygen, CPAP or intermittentpositive pressure ventilation were required, this was given witha Neopuff (Fisher & Paykel, Auckland, New Zealand) T-pieceresuscitation device. In the second time period, a small numberof infants were managed using a self-inflating resuscitationdevice (Laerdal, Stavanger, Norway).After resuscitation, data from the oximeter (HR, SpO2 and
signal quality) were downloaded to a computer using theNeO2m program16 (Dr Girvan Malcolm, Royal Prince AlfredHospital, Sydney, Australia). The data were analysed with Stata(Intercooled 10). We only analysed measurements where thesignal was considered normal, ie, no alarm messages (low IQsignal, low perfusion, sensor off, ambient light).These observational data are presented to illustrate the effects
on SpO2 and HR of resuscitation with either 100% oxygen or airwith backup 100% oxygen if the SpO2 was ,70% at 5 min. Thedata are presented as numbers and proportions (%) forcategorical variables, or means (SD) for normally distributedcontinuous variables and median (interquartile range) when the
distribution was skewed. SpO2 and HR during the first 10 minare illustrated by group using box plots showing the median,interquartile range (IQR) and range with outliers. We did notdefine primary or secondary outcomes a priori, thereforeinferential statistics have not been used to compare thesehistorical cohorts.
Figure 1 The Royal Women’s Hospitalguideline for managing administration andtitration of supplemental oxygen in thedelivery room. CPAP, continuous positiveairway pressure; FiO2, fractional inspiredoxygen; NICU, neonatal intensive careunit; SpO2, oxygen saturation.
Table 1 Characteristics of infants in the two groups
OX100
(n= 20)
OX21
(n= 105)
Gestational age (weeks)* 27 (1.6) 27 (1.6)
Birth weight (g)* 915 (300) 930 (293)
Male 13 (65) 67 (64)
Labour started 3 (15) 51 (48)
Full course of antenatalsteroids
18 (90) 87 (82)
Apgar score at 1 min 5 (5–7) 5 (3–7)
Apgar score at 5 min 8 (7–9) 8 (6–9)
Cord pH 7.28 (7.25–7.34) 7.3 (7.25–7.34)
Days in oxygen 3.5 (2–35.5) 20 (3–44)
Days treated with CPAP 9 (4–42) 9 (4–34)
Days treated withventilation
9 (3–18) 6 (3–12)
Died before discharge/transfer
3 (15) 12 (11)
Values are number (%) unless indicated otherwise.*Mean (SD).Median (interquartile range).pH available for eight infants in the OX100 group and 77 infants inthe OX21 group.CPAP, continuous positive airway pressure; OX100, received 100%oxygen; OX21, received 21% oxygen.
Original article
F88 Arch Dis Child Fetal Neonatal Ed 2009;94:F87–F91. doi:10.1136/adc.2008.141341
A29
In the OX100 group, verbal consent was obtained fromparents to monitor their infants in the DR. In the later group,parental consent was not obtained to monitor infants, asapplying an oximeter sensor for monitoring in the DR was thestandard of care for management of infants at ‘‘high risk’’ forreceiving active resuscitation in our institution.
RESULTSA total of 126 infants were studied. Pulse oximeter data wereavailable for 125 infants (sensor failure in one infant in the OX21
group). All 20 infants in the OX100 group received 100% oxygenbefore 1 min of age. In the OX21 group, 97/105 (92%) weresubsequently treated with supplemental oxygen at median(IQR) of 5.05 (4–5.5) min. Eight infants (8%) in the OX21 groupdid not receive supplemental oxygen in the DR. Tables 1 and 2present the clinical characteristics and DR interventions,respectively. No infants received external cardiac massage.
Changes in oxygen saturationFigure 2 shows the changes in SpO2 values for the two groupsover the first 10 min. The median SpO2 at 1 min was 60% forthe OX100 group and 55% for the OX21 group. By 2 min, theOX100 group had a median SpO2 of 84%, which continued to risesteadily to a median of 94% at 5 min and 96% by 10 min. Forthe OX21 group, the median SpO2 fell to 31% at 2 min, then roseto a median of 54% at 5 min, followed by a sharp rise to 81% at6 min, after supplementary 100% oxygen was started, reachinga median SpO2 of 91% at 10 min. After 5 min, the median SpO2
was very similar in the two groups. In the first 10 min afterbirth, 80% and 55% of infants in the OX100 and OX21 groups,respectively, had an SpO2 >95%.The eight infants not receiving supplemental oxygen were
similar in gestation to infants in both the OX100 and OX21
groups, with a mean (SD) gestational age of 27.5 (1) weeks;however, they were slightly larger with a mean (SD) birthweight of 1044 (167) g. Six received CPAP and three receivedintermittent positive pressure ventilation via a face mask. Themedian SpO2 in these eight infants at 1, 2, 5 and 10 min was60%, 71%, 87% and 93%, respectively.
Changes in HRFigure 3 shows the changes in HR over the first 10 min. Themedian HR in both groups at 1 min was ,100 beats/min: 76beats/min (OX100 group) and 93 beats/min (OX21 group). Inboth groups, the median HR increased to over 100 beats/min by
2 min, over 140 beats/min by 5 min and over 150 beats/min at10 min.
DISCUSSIONPulse oximetry is increasingly used during neonatal resuscita-tion17 18 with some centres using it to target a specific SpO2
range.19 Several studies report SpO2 changes in term or near-terminfants not requiring resuscitation in the first minutes afterbirth.13 14 20–22 Recent studies used pulse oximeters with algo-rithms that deal with low perfusion and motion artefact, bothof which are common in the DR.13 14 20–22 These studies reportedan SpO2 of ,60% at 1 min, with many infants taking at least10 min to achieve >90%. Data from our very preterm infantsinitially resuscitated with air and backup 100% oxygen had asimilar course. In our very preterm infants resuscitated with100% oxygen, the SpO2 rose more quickly.When the infants in our study were resuscitated initially with
air, the SpO2 levels were at the lower end of the normal range forhealthy term infants13 and rose into the ‘‘normal’’ range whenthey were treated with supplemental oxygen at about 5 min. By6 min, their median SpO2 was 81%. By 7 min, there was littledifference in the SpO2 between the OX21 group and the OX100
group. Fewer infants in the OX21 group had an SpO2 >95% inthe first 10 min than in the OX100 group.A systematic review found that air was more effective than
100% oxygen for resuscitation of asphyxiated term infants.23
SpO2 data available for some infants in these trials24–26 show nosignificant difference in SpO2 measurements for infants rando-mised to receive air or 100% oxygen. However, few very preterminfants were enrolled in these studies.Our hospital changed policy to starting resuscitation in air on
the basis of evidence from randomised trials comparing initialDR resuscitation with 100% oxygen or air plus 100% oxygen asneeded.8 9 26 When we developed our protocol, there were fewdata from randomised trials comparing 100% oxygen with anoxygen concentration other than 21% in preterm infants. Twostudies had randomised infants to either ,100% or .21%oxygen. Lundstrom et al27 randomised 70 infants ,33 weeks’gestation to receive air or 80% oxygen in the DR. Theyhypothesised that cerebral blood flow at 2 h of age might bereduced after a brief period of hyperoxia from 80% oxygen atbirth. They found that the cerebral blood flow was significantly(p,0.0001) higher in the group treated with air. They alsoshowed, in a subgroup of infants monitored with oximetry, thatthe mean SpO2 was significantly higher at 3, 5 and 7 min in the80% oxygen group than the air group. In the air group, 74% didnot receive supplemental oxygen. In those who received oxygen,the maximum was 50%. The reasons for the different oxygenrequirements from those in our study are that the infants in thestudy of Lundstrom et al were more mature and clinicalmethods were used rather than SpO2 to titrate oxygen in theair group. We have shown that clinicians’ ability to measurecolour28 or HR29 in the DR is weak. Harling et al30 randomised 63infants ,31 weeks’ gestation to either 50% or 100% oxygen inthe DR. They hypothesised that cytokine concentration inbronchoalveolar lavage fluid 12 h after birth would be highest ininfants treated with 100% oxygen, but found no significantdifference. In infants randomised to receive 50% oxygen, one-third had an FiO2 above 50% during resuscitation. They did notmeasure SpO2 in the DR.One approach to selecting the FiO2 to use in the DR is to use
SpO2 measurements to adjust the FiO2.31 In the neonatalintensive care unit, targeting a narrow range for SpO2 andavoiding hyperoxia is associated with reduced morbidity in
Table 2 Delivery room interventions
OX100
(n= 20)
OX21
(n= 105)
Nasopharyngeal suction 9 (45) 51 (48)
CPAP 16 (80) 72 (69)
IPPV 14 (70) 80 (76)
Endotracheal intubation 8 (40) 42 (40)
Surfactant administered 2 (10) 5 (5)
Oxygen administered 20 (100) 97 (92)
Time oxygen started (minfrom birth)*
1 (0.86–1.2) 5.05 (4–5.5)
Each infant may have received several interventions.Values are number (%) unless indicated otherwise.*Median (interquartile range).CPAP, continuous positive airway pressure; IPPV, intermittentpositive pressure ventilation.
Original article
Arch Dis Child Fetal Neonatal Ed 2009;94:F87–F91. doi:10.1136/adc.2008.141341 F89
A30
extremely-low-birthweight infants32 33 without a detrimentaleffect on developmental outcomes.32 It seems logical that a‘‘targeted oxygen delivery approach’’34 should be applied duringresuscitation. A small observational study35 titrated FiO2 againsttargeted SpO2 in 15 infants born at 24–29 weeks. They wereinitially resuscitated with 100% oxygen and the FiO2 adjusted tomaintain the SpO2 between 80% and 92%. The FiO2 was reducedfrom 100% to ,40%.
There are now three controlled studies on very preterminfants where the FiO2 has been titrated to the SpO2 after birth.Escrig et al36 randomised 28 infants ,29 weeks’ gestation toreceive 30% or 90% oxygen. The FiO2 was adjusted to achieve anSpO2 of 85%. By 5 min, the FiO2 was just above 50%, with nosignificant difference between the groups. Wang et al34
randomised infants ,32 weeks’ gestation to start resuscitationwith 100% oxygen or air. In the 100% oxygen group, the FiO2
was weaned if the SpO2 was .95% at 5 min. In the air group,the FiO2 was increased in 25% steps if the SpO2 was ,70% at3 min or 85% at 5 min, or to 100% if the HR was ,100 beats/min for 2 min or,60 beats/min for 30 s at any time. All infantsin the air group received oxygen from ,3 min. Infants in the100% oxygen group had significantly higher FiO2 from 1 to7 min, but from 8 to 20 min it was similar in the two groups.From 2 to 10 min, the SpO2 was higher in the group initiallyresuscitated with 100% oxygen.In the only randomised study to mask clinicians to the SpO2,
Rabi et al37 randomised 106 infants ,33 weeks’ gestation tothree groups. One received 100% oxygen throughout resuscita-tion, the second received an initial concentration of 100%,which could than be changed, and the third group started withair. In the last two groups, the FiO2 was changed by 20% every15 s until the SpO2 was between 85% and 92%. The mean timethat each group spent in the SpO2 target range was 11%, 21%and 16%, respectively (p,0.01). At the end of resuscitation, theFiO2 was similar in the two targeted groups.The safe SpO2 range for very preterm infants during
resuscitation is undefined. We targeted an SpO2 of 80–90%.
Escrig et al36 targeted 85%, Wang et al34 targeted 80–85% at5 min and 85–90% after 7 min, and Rabi et al37 targeted an SpO2
range of 85–92%. Each group used slightly different targetscombined with different resuscitation protocols, which willhave influenced the FiO2 used.Hyperoxia is common during resuscitation of preterm
infants. Tracy et al38 showed that, of 26 ventilated preterminfants of mean gestation 28 weeks (range 23–34), 38% werehyperoxic (PaO2 .100 mm Hg) 15 min after birth. Both Wanget al34 and our own study have shown that more of the infantswho started in 100% oxygen had an SpO2 of .95% than thosewho started in air. Rabi et al37 reported that infantsresuscitated in 100% oxygen spent 49% of the time abovethe SpO2 range of 85–92%. Therefore, in very preterm infants,hyperoxia appears to be a problem after starting resuscitationwith a high FiO2 and indicates that starting with a lower FiO2
may be preferable.Our results contribute to a growing body of evidence that it is
possible to adjust the FiO2 to keep SpO2 measurements within atargeted range during resuscitation. 34 36 37 Our study shows thatmost of our very preterm infants received supplemental oxygenwhen the initial resuscitation was with air and their SpO2 was,70% at 5 min or ,90% at 10 min. However, with differentSpO2 targets, the use of supplemental oxygen would bedifferent. We have shown that, using oximetry, it is possibleto titrate the FiO2 to the SpO2, with the SpO2 rising at a similarrate to that in term infants not receiving assistance in thedelivery room.13 14 21
When the initial resuscitation of very preterm infants hasbeen started with .90% oxygen, it has generally been possibleto reduce the FiO2 in response to the SpO2 to , 50% at 5 minand 35% at 10 min. In comparison, when the initial resuscita-tion gas was air or a low FiO2, the FiO2 at 5 min was similar towhen a high FiO2 was used. Therefore it now appears that,whatever the FiO2 used to initiate resuscitation, if the SpO2 ismonitored within a few minutes an appropriate FiO2 isachieved.
Figure 2 SpO2 shown at each minute, for the first 10 min after birth,from the group receiving 100% oxygen (OX100) and the group receiving21% oxygen (OX21). The box plots show the median, interquartile range,normal range and outliers. Small circles indicate outliers.
Figure 3 The heart rate shown at each minute, for the first 10 min afterbirth, for infants from the 100% oxygen (OX100) group and the 21% oxyen(OX21) group. The box plots show the median, interquartile range, normalrange and outliers. Small circles indicate outliers.
Original article
F90 Arch Dis Child Fetal Neonatal Ed 2009;94:F87–F91. doi:10.1136/adc.2008.141341
A31
HR is the most important indicator of an infant’s response toresuscitation.18 31 Using pulse oximetry in the DR providesclinicians with a continuous display of HR without having tointerrupt resuscitation to listen to the HR intermittently.Importantly we found that the HR increased after birth at asimilar rate in this group of very preterm infants, regardless ofwhether the infant was resuscitated with air or 100% oxygen.Our finding and those of other researchers34 36 have shown that,even with a low SpO2 during the first few minutes after birth,HR was similar to that achieved by healthy newborn terminfants not receiving assistance.13 14 21
Two important questions cannot be answered by our studybut could be addressed in future trials. Does pulse oximetry inthe DR improve outcomes for preterm infants? If pulseoximetry is effective, what is the safe SpO2 range for preterminfants in the first minutes after birth?
Acknowledgements: JAD and COFK are recipients of a RWH PostgraduateScholarship. ATP is the recipient of a Ter Meulen Fund grant for working visits, RoyalNetherlands Academy of Arts and Sciences, The Netherlands, and PGD is the recipientof a NHMRC Practitioner Fellowship. The work was supported by Australian NationalHealth and Medical Research Council Program Grant No 384100. We thank Dr GirvanMalcolm, Royal Prince Alfred Hospital, Sydney for his assistance with the NeO2Mprogram.
Competing interests: None.
Patient consent: Parental consent obtained.
REFERENCES1. O’Donnell CPF, Gibson AT, Davis PG. Pinching, electrocution, ravens’ beaks, and
positive pressure ventilation: a brief history of neonatal resuscitation. Arch Dis ChildFetal Neonatal Ed 2006;91:F369–73.
2. Saugstad OD. Optimal oxygenation at birth and in the neonatal period. Neonatology2007;91:319–22.
3. Kondo M, Itoh S, Isobe K, et al. Chemiluminescence because of the production ofreactive oxygen species in the lungs of newborn piglets during resuscitation periodsafter asphyxiation load. Pediatr Res 2000;47:524–7.
4. Kutzsche S, Ilves P, Kirkeby OJ, et al. Hydrogen peroxide production in leukocytesduring cerebral hypoxia and reoxygenation with 100% or 21% oxygen in newbornpiglets. Pediatr Res 2001;49:834–42.
5. Saugstad OD, Ramji S, Vento M. Oxygen for newborn resuscitation: how much isenough? Pediatrics 2006;118:789–92.
6. Vento M, Asensi M, Sastre J, et al. Resuscitation with room air instead of 100%oxygen prevents oxidative stress in moderately asphyxiated term neonates. Pediatrics2001;107:642–7.
7. Vento M, Asensi M, Sastre J, et al. Hyperoxemia caused by resuscitation with pureoxygen may alter intracellular redox status by increasing oxidized glutathione inasphyxiated newly born infants. Semin.Perinatol 2002;26:406–10.
8. Vento M, Asensi M, Sastre J, et al. Oxidative stress in asphyxiated term infantsresuscitated with 100% oxygen. J Pediatr 2003;142:240–6.
9. Vento M, Sastre J, Asensi MA, et al. Room-air resuscitation causes less damage toheart and kidney than 100% oxygen. Am J Respir.Crit Care Med 2005;172:1393–8.
10. Australian Resuscitation Council. Guideline 13.1 Introduction to resuscitation ofthe newborn infant. 2006. http://www.resus.org.au/ (accessed 17 Oct 2008).
11. Canadian Paediatric Society. Recommendations for specific treatmentmodifications in the Canadian context: addendum to the 2006 NRP Provider Textbook.Canadian Paediatric Society 2006. 9-9-0007. http://www.cps.ca/English/ProEdu/NRP/addendum.pdf (accessed 23 Oct 2008).
12. Saugstad OD. Oxidative stress in the newborn: a 30-year perspective. Biol Neonate2005;88:228–36.
13. Kamlin CO, O’Donnell CPF, Davis PG, et al. Oxygen saturation in healthy infantsimmediately after birth. J Pediatr 2006;148:585–9.
14. Rabi Y, Yee W, Chen SY, et al. Oxygen saturation trends immediately after birth.J Pediatr 2006;148:590–4.
15. O’Donnell CPF, Kamlin CO, Davis PG, et al. Feasibility of and delay in obtaining pulseoximetry during neonatal resuscitation. J Pediatr 2005;147:698–9.
16. Malcolm G. Neonatal oxygen saturation download and analysis - user manual. 2007.http://www.cs.nsw.gov.au/rpa/ (accessed 23 Oct 2008).
17. O’Donnell CPF, Davis PG, Morley CJ. Use of supplementary equipment forresuscitation of newborn infants at tertiary perinatal centres in Australia and NewZealand. Acta Paediatr 2005;94:1261–5.
18. Leone TA, Rich W, Finer NN. A survey of delivery room resuscitation practices in theUnited States. Pediatrics 2006;117:164–75.
19. Finer NN. Rich WD. Neonatal resuscitation: raising the bar. Curr Opin Pediatr2004;16:157–62.
20. Toth B, Becker A, Seelbach-Gobel B. Oxygen saturation in healthy newborn infantsimmediately after birth measured by pulse oximetry. Arch Gynecol Obstet2002;266:105–7.
21. Altuncu E, Ozek E, Bilgen H, et al. Percentiles of oxygen saturations in healthy termnewborns in the first minutes of life. Eur J Pediatr 2008 ;167 :687–8.
22. Mariani G, Dik PB, Ezquer A, et al. Pre-ductal and post-ductal O2 saturation inhealthy term neonates after birth. J Pediatr 2007;150:418–21.
23. Davis PG, Tan A, O’Donnell CPF, et al. Resuscitation of newborn infants with 100%oxygen or air: a systematic review and meta-analysis. Lancet 2004;364:1329–33.
24. Ramji S, Ahuja S, Thirupuram S, et al. Resuscitation of asphyxic newborn infantswith room air or 100% oxygen. Pediatr Res 1993;34:809–12.
25. Rao R,.Ramji S. Pulse oximetry in asphyxiated newborns in the delivery room. IndianPediatr 2001;38:762–6.
26. Saugstad OD, Rootwelt T, Aalen O. Resuscitation of asphyxiated newborn infantswith room air or oxygen: an international controlled trial: the Resair 2 study. Pediatrics1998;102:e1.
27. Lundstrom KE, Pryds O, Greisen G. Oxygen at birth and prolonged cerebralvasoconstriction in preterm infants. Arch Dis Child Fetal Neonatal Ed 1995;73:F81–6.
28. O’Donnell CPF, Kamlin CO, Davis PG, et al. Clinical assessment of infant colour atdelivery. Arch Dis Child Fetal Neonatal Ed 2007;92:F465–7.
29. Kamlin COF, O’Donnell CPF, Everest N, et al. Accuracy of clinical assessment ofinfant heart rate in the delivery room. Resuscitation 2006;71:319–21.
30. Harling AE, Beresford MW, Vince GS, et al. Does the use of 50% oxygen at birth inpreterm infants reduce lung injury? Arch Dis Child Fetal Neonatal Ed 2005;90:F401–5.
31. American Heart Association. 2005 American Heart Association (AHA) Guidelinesfor cardiopulmonary resuscitation (CPR) and Emergency cardiovascular care (ECC) ofpediatric and neonatal patients: neonatal resuscitation guidelines. Pediatrics2006;117:1029–38.
32. Deulofeut R, Critz A, Adams-Chapman I, et al. Avoiding hyperoxia in infants(1250 g is associated with improved short- and long-term outcomes. J Perinatol2006;26:700–5.
33. Chow LC, Wright KW, Sola A. Can changes in clinical practice decrease theincidence of severe retinopathy of prematurity in very low birth weight infants?Pediatrics 2003;111:339–45.
34. Wang CL, Anderson C, Leone TA, et al. Resuscitation of preterm neonates by usingroom air or 100% oxygen. Pediatrics 2008;121:1083–9.
35. Kopotic RJ, Lindner W. Assessing high-risk infants in the delivery room with pulseoximetry. Anesth.Analg 2002;94:S31–6.
36. Escrig R, Arruza L, Izquierdo I, et al. Achievement of targeted saturation values inextremely low gestational age neonates, resuscitated with low or high oxygenconcentrations: a prospective randomized trial. Pediatrics 2008;121:875–81.
37. Rabi Y, Nettel-Aguirre A, Singhal N. Room air versus oxygen administration duringresuscitation of preterm infants (ROAR Study). Meeting of Pediatric AcademicSocieties, Hawaii, 2008. E-PAS2008: 5127.5
38. Tracy M, Downe L, Holberton J. How safe is intermittent positive pressureventilation in preterm babies ventilated from delivery to newborn intensive care unit?Arch Dis Child Fetal Neonatal Ed 2004;89:F84–7.
Original article
Arch Dis Child Fetal Neonatal Ed 2009;94:F87–F91. doi:10.1136/adc.2008.141341 F91
A32
ORIGINAL ARTICLE
Neonatal resuscitation 1: a model to measureinspired and expired tidal volumes and assessleakage at the face maskC P F O’Donnell, C O F Kamlin, P G Davis, C J Morley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
See end of article forauthors’ affiliations. . . . . . . . . . . . . . . . . . . . . . .
Correspondence to:Dr O’Donnell, RoyalWomen’s HospitalMelbourne, 132 GrattanStreet, Carlton, Victoria3053, Australia;[email protected]
Accepted 18 March 2005Published Online First4 May 2005. . . . . . . . . . . . . . . . . . . . . . .
Arch Dis Child Fetal Neonatal Ed 2005;90:F388–F391. doi: 10.1136/adc.2004.064683
Background: Neonatal resuscitation is a common and important intervention, and adequate ventilation isthe key to success. In the delivery room, positive pressure ventilation is given with manual ventilationdevices using face masks. Mannequins are widely used to teach and practise this technique. During bothsimulated and real neonatal resuscitation, chest excursion is used to assess tidal volume delivery, andleakage from the mask is not measured.Objective: To describe a system that allows measurement of mask leakage and estimation of tidal volumedelivery.Methods: Respiratory function monitors, a modified resuscitation mannequin, and a computer were usedto measure leakage from the mask and to assess tidal volume delivery in a model of neonatal resuscitation.Results: The volume of gas passing through a flow sensor was measured at the face mask. This was a goodestimate of the tidal volume entering and leaving the lung in this model. Gas leakage between the maskand mannequin was also measured. This occurred principally during inflation, although gas leakageduring deflation was seen when the total leakage was large. A volume of gas that distended the mask butdid not enter the lung was also measured.Conclusion: This system can be used to assess the effectiveness of positive pressure ventilation given usinga face mask during simulated neonatal resuscitation. It could be useful for teaching neonatal resuscitationand assessing ventilation through a face mask.
Internationally agreed consensus statements on neonatalresuscitation advise that the key to success is adequateventilation.1 2 It is recommended that breathing be assisted
by giving positive pressure ventilation (PPV) with a manualventilation device via a face mask.1 2 Observation of chestwall movement is recommended to assess the adequacy ofventilation—that is, gas delivery to the lungs1 2—althoughthe precision and accuracy of this assessment is not known.The technique of ‘‘bag and mask’’ ventilation is generallytaught and practised using mannequins. Teaching andassessment of these techniques using mannequins form anintegral part of neonatal life support courses, which are amandatory part of physician, midwife, and neonatal nursetraining in many countries.3 4
According to consensus statements, most newborn infantscan be adequately ventilated with a bag and mask.1 2
However, the evidence to support this statement is lacking,the few studies of infants in the delivery room reporting thattidal volumes adequate for gas exchange were rarelydelivered through a mask.5–7 There are no reported studiesinvestigating resuscitation of very preterm infants using amask. One study has assessed face masks used for neonatalresuscitation.8 The participants in that study had little or noexperience of resuscitating infants, manual ventilationdevices were not used, and leakage was not measured.Giving PPV to a mannequin through a face mask mimics
practice in the delivery room in that ‘‘chest’’ excursion is usedto assess the adequacy of ventilation, and there is noindication of the extent of the leak between the mask andface. To assess the adequacy of ventilation during resuscita-tion, measurement of the volume of gas entering and leavingthe lungs (the tidal volume, VT) would be ideal. This is,however, impractical unless the infant is intubated. In theabsence of a direct measurement, an estimate of the tidal
volume delivered during bag and mask ventilation, whichwas more accurate than the current ‘‘gold standard’’ of chestexcursion, would be useful. A measure of leakage occurringat the face mask would also be useful, as large leaks willreduce the tidal volumes delivered and thus the effectivenessof PPV.We therefore modified a resuscitation mannequin so that
we could measure the volume of gas passing through themask and the tidal volume delivered to a test lung duringsimulated neonatal resuscitation. We wished to calculate thevolume of gas leaking at the mask and determine how wellthe tidal volume returning through the mask represented thetidal volume returning from the lung.
MATERIALS AND METHODSManual ventilation devices and face masksThe manual ventilation devices used at our hospital are theNeopuff infant resuscitator (Fisher & Paykel Healthcare,Auckland, New Zealand) and the Laerdal infant resuscitator(Laerdal Medical, Oakleigh, Australia). The Neopuff is a Tpiece device that requires a compressed gas source. It has avalve on the outlet which allows a positive end expiratorypressure to be set for a given flow rate. Occlusion of this valvegenerates a predetermined peak inspiratory pressure. TheLaerdal Infant Resuscitator is a 240 ml silicone self inflatingbag. We routinely use round silicone Laerdal face masks(Laerdal Medical) with both devices at our hospital, and usedthe appropriate size mask (0/1) for the mannequin in thisstudy.
Abbreviations: PPV, positive pressure ventilation; VMASK, volume of gasthat, under pressure, distends the mask but does not enter the lung; VT,tidal volume; VTE(mask), expiratory tidal volume at the mask; VTI(mask),inspiratory tidal volume at the mask
F388
www.archdischild.com
A33
Modification of the mannequinThe Laerdal Resusci Baby mannequin (Laerdal Medical) isthe most widely used resuscitation mannequin and has beenfound to be the most realistic for the purposes of simulatedPPV.9 This mannequin is supplied with a bag in its ‘‘thorax’’.When PPV is given appropriately, this bag inflates and slowlydeflates, causing visible ‘‘chest’’ rise and fall. We replaced thebag with a test lung with a baseline volume of 50 ml (Drager,Lubeck, Germany) and connected it via an airtight seal to themannequin’s ‘‘oropharynx’’, so that its inflation and defla-tion caused ‘‘chest’’ excursion similar to that of the originalmannequin. A pressure monitoring line was connectedimmediately proximal to the test lung. The compliance ofthis model, calculated by measuring the inspired volume ofthe whole system when pressurised to 25 cm H2O, was
0.46 m/cm H2O, comparable to that of an infant withrespiratory distress syndrome.10
Respiratory monitorTwo Florian Respiratory Monitors (Acutronic MedicalSystems, Zug, Switzerland) were used to measure gas flow.This monitor uses a flow sensor with a hot wire anemometerwith minimal (,1 ml) dead space to detect gas flow. Themonitor calculates the volumes of gas passing through thesensor by integration of the flow signal. The flow sensorsfrom the two monitors were placed in series, and theirvolume measurement calibrated simultaneously using a fixedvolume syringe. The monitors measure airway pressuresdirectly and were calibrated against a column of water. Theoutput from the Florian monitors was acquired using ananalogue-digital converter using the Spectra software pro-gram (Grove Medical, London, UK). This is a computerprogram specifically designed for the acquisition and analysisof respiratory signals.
Values measuredWe placed one flow sensor (FS1, fig 1) between the manualventilation device and the face mask. With this sensor, wemeasured the volume of gas passing from the device throughthe mask—the inspiratory tidal volume at the mask(VTI(mask))—and the volume of gas returning from themannequin through the mask—the expiratory tidal volumeat the mask (VTE(mask)). We placed the second sensor (FS2,fig 1) in the mannequin’s ‘‘airway’’, proximal to the test lung.With this sensor we measured the tidal volume (VT) enteringand leaving the test lung.
Values calculatedIf the volume of gas passing from the device through themask exceeded that returning from the mannequin throughthe mask, there was a leak between the mask and the face.This was calculated as their difference expressed as apercentage of the inspiratory tidal volume at the mask:leak (%) = ((VTI(mask) 2 VTE(mask))/VTI(mask))6 100.If leakage from the mask was detected, we wished
to determine whether it occurred during inspiration orexpiration.
Flow sensor 2(FS2)Mannequin
Face mask
Leak duringinflation
Leakduringdeflation
Manual ventilationdevice
VT
VTE(mask)VTI(mask)
Flow sensor (FS1)
Pressuremonitoringline
Test lung
Figure 1 Diagram of apparatus used to assess accuracy of measurements made with flow sensor 1 (FS1) using flow sensor 2 (FS2). VTI(mask),inspiratory tidal volume at the mask; VTE(mask), expiratory tidal volume at the mask; VT, tidal volume.
16
12
8
0
4
Volu
me
at lu
ng F
S2 (m
l)
16
12
8
0
4
Volu
me
at m
ask
FS1
(ml)
Inflation Deflation
Figure 2 Trace obtained when the face mask was applied firmly to aflat bench top so that no gas escaped, and positive pressure inflation wasapplied. This illustrates a volume of gas distending the mask that appearsto contribute to tidal volume, but never enters the lung. FS1, Flow sensor1; FS2, flow sensor 2.
Neonatal resuscitation: tidal volume and mask leakage F389
www.archdischild.com
A34
We hypothesised that if the volume of gas passing from thedevice through the mask (VTI(mask)) did not equal the volumereturning through the mask from the mannequin (VTE(mask))or the tidal volume measured in the mannequin (VT), therecould be three potential reasons:
(1) Leakage during inflation. In the presence of a poor sealbetween the mask and face, some of the gas passingthrough FS1 would escape during inflation and not passthrough FS2 at the lung. The volume leaking duringinflation would thus contribute to VTI(mask) but not to VT
or VTE(mask).
(2) Leakage during deflation. In the presence of a poor sealbetween the mask and face, some of the gas leaving thelung through FS2 would escape around the mask duringdeflation and not pass through FS1, resulting in VTE(mask)
underestimating VT.
(3) Distension of the mask. The mask used is made ofdistensible silicone; thus we hypothesised that, in thepresence of a reasonable seal between the mask and face,gas under pressure would distend the mask but not enterthe lung (VMASK). This would pass through FS1 duringinflation and back through FS1 during deflation, thuscontributing to VTI(mask) and VTE(mask) but not to VT. Thisvolume was demonstrated by placing the mask flat on abench top and giving positive pressure inflations (fig 2).
To test this system, we recorded 100 inflations, applied by aconsultant neonatologist, with various amounts of leakage atthe face mask, with each manual ventilation device.Inflations with a peak inspiratory pressure of 25 cm H2Oand positive end expiratory pressure 5 cm H2O were givenwith the Neopuff. With the Laerdal bag a manometer wasplaced in the circuit and a peak inspiratory pressure of25 cm H2O was targeted for inflations. Data were analysedusing SPSS for Windows (SPSS Inc, Chicago Illinois, USA).Results are expressed as mean (SD).
RESULTSThe mannequin’s airway and lung were leak free, with adifference between the volume entering and leaving the lungof –0.2 (0.5)%. There were no significant differences in theVTI(mask), VTE(mask), or VT delivered with each manualventilation device.
Relationship of VTE(mask) to VT
The volume returning from the mannequin through the mask(VTE(mask)) was a good estimate of the volume leaving thelung (VT) (fig 3). When there were large leaks (.51%),VTE(mask) tended to underestimate VT because some of the gasreturning from the lung escaped around the mask duringdeflation (see under Leak). With smaller leaks, VTE(mask)
tended to overestimate VT because a volume of gas distendedthe mask but did not enter the lung (see under Volumedistending the mask and not entering the lung (VMASK)).Overall, VTE(mask) was 100.6 (24.8)% of the lung VT.
LeakIf the volume of gas passing from the ventilation devicethrough the mask was greater than that returning from themannequin through the mask—that is, VTI(mask) .VTE(mask)—there was a leak between the mask and the face.When the volume returning from mannequin through themask (VTE(mask)) was greater than or equal to the volumeleaving the lung (VT), all of the leak occurred duringinflation.When the volume leaving the lung (VT) was greater than
the volume returning from the mannequin through the mask(VTE(mask)), a proportion of the leak occurred in deflation.Leaks during deflation were never seen without a leak duringinflation and only seen when total leakage was greater than51%. When there was a large leak, the proportion occurringduring inflation (((VTI(mask) 2 VT)/( VTI(mask) 2 VTE(mask)))6100) was 89 (10)%, and the proportion during deflation (((VT
2 VTE(mask))/(VTI(mask) 2 VTE(mask)))6 100) was 11 (10)%.
Volume distending the mask and not entering the lung(VMASK)When the volume passing from the device through the mask(VTI(mask)) equalled the volume returning from the lung(VTE(mask)), there was no leakage from the mask. Duringthese inflations, VMASK was readily demonstrated, as thevolume passing through FS1 was greater than that passingthrough FS2 and entering the lung. Although this becameprogressively smaller as the total leak increased, VMASK wasseen with leaks up to 51%. This volume was 2.4 (1.4) ml,which represented 18.3 (10.5)% of VT.
DISCUSSIONPositive pressure ventilation using manual ventilation devicesvia a face mask is an important skill. It is widely taught andpractised using mannequins. We have developed a systemthat allows estimation of tidal volume delivered andmeasurement of mask leak during simulated PPV. Thepressures used may also be recorded using this system. Thissystem was developed with a view to recording realresuscitations in the delivery room.Studies of PPV through a mask during resuscitation in
the delivery room are few. In a study of term infants, ‘‘bagand mask’’ ventilation seemed relatively inefficient, withtidal exchange substantially less than that seen afterintubation and rarely sufficient to produce adequate alveolarventilation.5 It is worth noting that the apparatus used tomeasure expired tidal volumes in this5 and other assessmentsof bag and mask ventilation6 7 was similar to ours, in that theflow sensor (pneumotachograph) was placed between themanual ventilation device and mask. Although the maskused for these studies differed from ours and its propertiesare probably not identical, there was leakage from this maskduring inflation. It is thus reasonable to speculate that theremay also have been leakage during expiration. It is unclearhow the expired tidal volume reported (equivalent to ourVTE(mask)) related to the expired tidal volume leaving thelungs.
22
18
14
16
12
6
8
4
2
0
10
222016 181410 12
Tidal volume at lung (ml)
R2 = 0.70Expi
red
volu
me
at m
ask
(ml)
840 2
20
6
Figure 3 Scatter plot showing the relation of the expired tidal volume atthe mask to the tidal volume at the lung.
F390 O’Donnell, Kamlin, Davis, et al
www.archdischild.com
A35
It remains unclear how effective bag and mask ventilationin the delivery room is as a means of delivering a satisfactorytidal volume, particularly in preterm infants, for which thereare no reported studies. Moreover, this technique is notwithout its complications; bradycardia caused by applicationof a mask during neonatal resuscitation has beenreported.11 12 Detailed studies of the efficacy of PPV througha face mask in newborn infants are needed, particularly inpreterm infants. Applying a flow sensor between the manualventilation device and face mask in the delivery room maygive valuable insights into the effect of PPV on newborninfants during resuscitation.
CONCLUSIONSWe have developed a system for practising and assessing PPVusing a face mask. In this model, the volume of gas returningthrough the face mask from the mannequin is a goodestimate of the tidal volume entering and leaving the lung.Most leakage from face masks occurs during inflation,although a small proportion does occur during deflation
when the total leakage is large (.51%). With leaks of up to51%, a volume of gas distends the mask but does not enterthe lungs. These data will be useful for further evaluation ofresuscitation equipment and techniques in bench top studiesand in the delivery room.
ACKNOWLEDGEMENTSCPFO’D is in receipt of the Royal Women’s Hospital PostgraduateDegree Scholarship. PGD is supported by a National Health andMedical Research Council Fellowship. We thank Dr Brendan Murphyand an anonymous reviewer for helpful comments that haveconsiderably improved this paper.
Authors’ affiliations. . . . . . . . . . . . . . . . . . . . .
C P F O’Donnell, C O F Kamlin, P G Davis, C J Morley, Royal Women’sHospital Melbourne, Victoria 3053, Australia
Competing interests: none declared
REFERENCES1 Kattwinkel J, Niermeyer S, Nadkarni V, et al. An Advisory Statement From
the Pediatric Working Group of the International Liaison Committee onResuscitation. Pediatrics 1999;103:e56.
2 Contributors and Reviewers for the Neonatal Resuscitation Guidelines.International Guidelines for Neonatal Resuscitation: an excerpt from theGuidelines 2000 for Cardiopulmonary Resuscitation and EmergencyCardiovascular Care: International Consensus on Science. Pediatrics2000;106:e29.
3 Kattwinkel J, ed. Neonatal resuscitation program: textbook of neonatalresuscitation, 4th ed. Washington DC: American Academy of Pediatrics/American Heart Association, 2000.
4 Richmond S, ed. Resuscitation at birth: the newborn life support providermanual. London: Newborn Life Support Working Party, Resuscitation Council(UK), 2001.
5 Milner AD, Vyas H, Hopkin IE. Efficacy of facemask resuscitation at birth. BMJ1984;289:1563–5.
6 Field D, Milner AD, Hopkin IE. Efficiency of manual resucitators at birth. ArchDis Child 1986;61:300–2.
7 Hoskyns EW, Milner AD, Hopkin IE. A simple method of face maskresuscitation at birth. Arch Dis Child 1987;62:376–8.
8 Palme C, Nystrom B, Tunell R. An evaluation of the efficiency of face masks inthe resuscitation of newborn infants. Lancet 1985;1:207–10.
9 Howells R, Madar J. Newborn resuscitation training: which manikin.Resuscitation 2002;54:175–81.
10 Morley CJ, Greenough A. Respiratory compliance in premature babiestreated with artificial surfactant (ALEC). Arch Dis Child 1991;66:467–71.
11 Finer NN, Rich W. Neonatal resuscitation: toward improved performance.Resuscitation 2002;53:47–51.
12 Sinclair JC, Bracken MB, eds. Effective care of the newborn infant. Oxford:Oxford University Press, 1999:30.
What this study adds
N We have adapted the most commonly used mannequinto allow measurement of leakage from masks andestimation of tidal volume delivered to the lungs
N The volume of gas returning through the mask is agood estimate of the volume leaving the lungs in thismodel
What is already known on this topic
N Resuscitation mannequins are used to teach andpractise ‘‘bag and mask’’ ventilation techniques
N During mask ventilation in both simulated and realneonatal resuscitation, there is no indication of theextent of leakage from the mask, and chest wallmovement is used to assess the adequacy of ventilation
Neonatal resuscitation: tidal volume and mask leakage F391
www.archdischild.com
A36
Assessing the effectiveness of two round neonatalresuscitation masks: study 1Fiona E Wood,1 Colin J Morley,1,2,3,4 Jennifer A Dawson,1 C Omar F Kamlin,1
Louise S Owen,1 Susan Donath,3,4 Peter G Davis1,2,3,4
1 Division of Neonatal Services,The Royal Women’s Hospital,Carlton, Victoria 3053, Australia;2 Departments of Obstetric andGynaecology, University ofMelbourne, Victoria 3053,Australia; 3 Murdoch Children’sResearch Institute, MelbourneVictoria 3052, Australia;4 Department of Paediatrics,University of Melbourne, Victoria3052, Australia
Correspondence to:Professor Colin J Morley,Division of Neonatal Services,The Royal Women’s Hospital,132 Grattan Street, Carlton,Victoria, 3053, Australia; [email protected]
Accepted 14 November 2007Published Online First26 November 2007
ABSTRACTBackground: Positive pressure ventilation (PPV) via aface mask is an important skill taught using manikins.There have been few attempts to assess the effective-ness of different face mask designs.Aim: To determine whether leak at the face mask duringsimulated neonatal resuscitation differed between a newround mask design and the current most widely usedmodel.Method: 50 participants gave PPV to a modified manikindesigned to measure leak at the face mask. Leak wascalculated from the difference between the inspired andexpired tidal volumes.Results: Mask leak varied widely with no significantdifference between devices; mean (SD) percentage leakfor the Laerdal round mask was 55% (31) and with theFisher & Paykel mask it was 57% (25).Conclusion: We compared a new neonatal face maskwith an established design and found no difference inleak. On average the mask leak was.50% irrespective ofoperator experience or technique.
Studies of manikins and infants reveal large leaksfrom face masks.1–3 Available masks are eitherround or anatomically shaped.4 Cushioned rimmedmasks were advocated,2 however the siliconeLaerdal ‘‘round’’ (LR) mask (Laerdal, Stavanger,Norway) is the most widely used.1 2 4 A new roundmask (Fisher & Paykel Healthcare, Auckland, NewZealand) is claimed to be more effective. Weinvestigated whether leak between the mask andface during simulated neonatal resuscitation dif-fered between these two masks.
METHODOLOGY
Setting and participantsNeonatal medical, nursing staff and midwives ofthe Royal Women’s Hospital, Melbourne, Australiaparticipated. All had received training in neonatalresuscitation.
Face masksThe LR mask is the only design in routine use atour hospital. The size 0/1 Laerdal mask and theFisher & Paykel (FP) 60 mm round mask were theappropriate sizes for the manikin (fig 1). Noparticipant had previously used the FP mask.
Manual ventilation deviceThe Neopuff Infant Resuscitator (Fisher & PaykelHealthcare, Auckland, New Zealand) is a pressurelimited T-piece device where the peak inspiratorypressure (PIP) and positive end expiratory pressure
(PEEP) are set and displayed. All participants wereexperienced in its use. The hospital protocolsuggested a gas flow of 8 l/min, a PEEP of 5 cmH2O and a PIP of 30 cm H2O.
Modified manikinA Laerdal Resusci baby manikin (Laerdal,Stavanger, Norway) was modified by removingthe lung and stomach bags, and positioning a50 ml test lung (Drager, Lubeck, Germany) intothe chest so chest excursion mimicked that of anunaltered training manikin.2 5 The test lung wasconnected by non-distensible tubing to the mouthwith an airtight seal. A pressure monitoring linewas connected to the airway. The system com-pliance when pressurised to 30 cm H2O was0.5 ml/cm H2O, with a maximal lung volume of65 ml.
Respiratory monitor, recording equipment andvalues measuredA Florian respiratory monitor (Acutronic MedicalSystems, Ag, Switzerland) measured inflatingpressures and gas flow through the mask. Airwaypressure was calibrated against a column of water;a 10 ml syringe was used to calibrate volumemeasurement. The flow sensor was placedbetween the T-piece and mask; tidal volume wascalculated by integration of the flow signal.Percentage leak at the face mask was the differencebetween the inspired and expired tidal volumes,expressed as a percentage of the inspired tidalvolume. O’Donnell et al showed this systemprovides a good estimate of the tidal volumes andleak.5
Spectra software (Grove Medical, London, UK)was used to acquire the output from the Florianmonitor through an A:D converter to a computer.Neither screen was visible to participants.
Study protocolParticipants ventilated the manikin for 2 minutesat a rate of 40–60 breaths per minute according tothe hospital protocol, ensuring chest rise. Theyknew we were recording tidal volume, mask leakand inflation pressures.The mean ventilation rate, PIP, PEEP, inspired
and expired tidal volumes and percentage leak weremeasured for a 2 minute period of ventilation witheach mask, excluding the first five inflations.
Observations and participant surveyMask placement and hold techniques, includingthe use of jaw lift were recorded. Participants were
Short report
Arch Dis Child Fetal Neonatal Ed 2008;93:F235–F237. doi:10.1136/adc.2007.117713 F235
A37
asked ‘‘Do you think your average mask leak was large orsmall?’’ (five-point Likert scale).
Randomisation, primary outcome and power calculationThe order of assessment of the masks was randomised. Thesample size was calculated using the mean percentage leak of70% (SD 30%) for the Laerdal mask.2 To detect a 15% differencein mean leak with an alpha value of 0.05 and power of 80%, 34participants were required.
Statistical analysisData were analysed using SPSS. A paired t test compared leakbetween the two masks, p,0.05 was considered significant.One-way ANOVA assessed the influence of placement-and-holdtechnique on mask leak.
RESULTS
Participants and years of experienceThere were 10 participants in each of the professionalcategories. The median years of experience were: consultants18.5 years, fellows (SpR grade) 3.3 years, registrars (SHO grade)0.1 years, neonatal nurses 9.5 years and midwives 3.5 years. Atotal of 10 084 inflations were recorded.
Techniques of mask application and holds used to form a sealwith the faceParticipants held the masks in four ways: (1) stem hold, thethumb and index finger gripping the stem; (2) two-point tophold, applying pressure to two points of the top flat portion ofthe mask; (3) rim hold, encircling much of the rim of the upperflat portion of the mask; (4) any other hold.
LR maskFrequencies were: stem hold 24%, two-point top hold 46%, rimhold 28% and other 2%. The mask was rolled onto the face by14% and the remaining 86% placed the mask directly onto theface. Jaw lift was applied by 50%.
Figure 1 Photograph showing the Laerdal round mask (left) and thenew Fisher & Paykel neonatal resuscitation mask (right).
Figure 2 Percentage leak at the face mask for each mask design (LRdark shading, FP light shading) and professional group. Box-plots showmedian values (solid lines), interquartile range (margins of box), range ofdata and any outliers (circles).
Figure 3 Participant assessment of facemask leak of the Laerdal round mask (left)and the Fisher & Paykel mask (right)compared with actual leak. Box-plotsshow median values (solid black lines),interquartile range (margins of box), rangeof data and any outliers (circle).
Short report
F236 Arch Dis Child Fetal Neonatal Ed 2008;93:F235–F237. doi:10.1136/adc.2007.117713
A38
FP maskFrequencies were: stem hold 26%, two-point top hold 44%, rimhold 28% and other 2%. The mask was rolled onto the face by10%, with 90% directly placing the mask. Jaw lift was appliedby 58%.
PIP, PEEP and ventilation rateThere were no significant differences in PIP, PEEP andventilation rates between the two masks.
Percentage leak at the face maskThe mean (SD) leak for the LR mask was 55% (31) and FP maskwas 57% (25), there was no significant difference between thedevices.Figure 2 shows a box-plot where percentage leak is grouped
by professional category for each mask. A large variation in leakwas found for all categories. One-way ANOVA analysis showedno significant difference in leak for either mask when groupedby mask hold, placement technique or use of jaw lift.
Participant assessment of mask leak compared with actual leakFigure 3 shows box-plots for LR and FP percentage leak groupedby self assessed leak for each mask. There was wide variation ofactual leak in nearly all of the self-assessed leak categories forboth masks. No one thought they had a very large leak with theLR mask, despite many with leaks of 80 to 100%. For eithermask, no one who said they had a very small/nil leak had a leakless than 20%.
DISCUSSIONWe found mask leak varied widely with both masks fromalmost zero to 100% with no significant difference between thetwo masks.Palme et al1 studied face mask efficiency by measuring leak
indirectly. They reported that the Laerdal mask ‘‘leaked’’ least,suggesting that the other designs must have had considerableleaks.
The techniques of holding the mask on the face, the way themask was placed on the manikin’s face and the use of jaw liftshowed considerable variation, with three distinct hold techni-ques being used. The technique used made no significantdifference to the mask leak.The set PIP could be achieved despite very large mean leaks
(data not shown). This is consistent with findings by O’Donnellet al.2 6 Overall, the participants were unable to accurately self-assess their leak.
CONCLUSIONWe compared a new mask with an established design and foundno difference in leak. On average the mask leak was .50%irrespective of operator experience or technique. Many opera-tors were unaware of the magnitude of the leak.Further studies of mask techniques, particularly directed
towards detecting and reducing leak, are required to improveeffectiveness of PPV.
Acknowledgements: We thank the staff of the Royal Women’s Hospital Melbournefor their participation, and Fisher & Paykel Healthcare for providing samples of theirneonatal resuscitation mask. Supported in part by a RWH Postgraduate degreescholarship (COFK) and a NHMRC Practitioner Fellowship (PGD).
Funding: None.
Competing interests: None.
REFERENCES1. Palme C, Nystrom B, Tunell R. An evaluation of the efficiency of face masks in the
resuscitation of newborn infants. Lancet 1985;1:207–10.2. O’Donnell CP, Davis PG, Lau R, et al. Neonatal resuscitation 2: an evaluation of
manual ventilation devices and face masks. Arch Dis Child Fetal Neonatal Ed2005;90:F392–6.
3. Milner AD, Vyas H, Hopkin IE. Efficacy of facemask resuscitation at birth. BMJ (ClinRes Ed) 1984;289:1563–5.
4. O’Donnell CP, Davis PG, Morley CJ. Positive pressure ventilation at neonatalresuscitation: review of equipment and international survey of practice. Acta Paediatr2004;93:583–8.
5. O’Donnell CP, Kamlin CO, Davis PG, et al. Neonatal resuscitation 1: a model tomeasure inspired and expired tidal volumes and assess leakage at the face mask. ArchDis Child Fetal Neonatal Ed 2005;90:F388–91.
6. O’Donnell CP, Davis PG, Lau R, et al. Neonatal resuscitation 3: manometer use in amodel of face mask ventilation. Arch Dis Child Fetal Neonatal Ed 2005;90:F397–400.
Short report
Arch Dis Child Fetal Neonatal Ed 2008;93:F235–F237. doi:10.1136/adc.2007.117713 F237
A39
Improved techniques reduce face mask leak duringsimulated neonatal resuscitation: study 2Fiona E Wood,1 Colin J Morley,1,2,3,4 Jennifer A Dawson,1 C Omar F Kamlin,1
Louise S Owen,1 Susan Donath,3,4 Peter G Davis1,2,3,4
1 Division of Neonatal Services,The Royal Women’s Hospital,Carlton, Victoria 3053, Australia;2 Departments of Obstetric andGynaecology, University ofMelbourne, Victoria 3053,Australia; 3 Murdoch Children’sResearch Institute, MelbourneVictoria 3052, Australia;4 Department of Paediatrics,University of Melbourne, Victoria3052, Australia
Correspondence to:Professor Colin J Morley,Division of Neonatal Services,The Royal Women’s Hospital,132 Grattan Street, Carlton,Victoria, 3053, Australia; [email protected]
Accepted 14 November 2007Published Online First26 November 2007
ABSTRACTBackground: Techniques of positioning and holdingneonatal face masks vary. Studies have shown that leakat the face mask is common and often substantialirrespective of operator experience.Aims: (1) To identify a technique for face maskplacement and hold which will minimise mask leak. (2) Toinvestigate the effect of written instruction and demon-stration of the identified technique on mask leak for tworound face masks.Method: Three experienced neonatologists comparedmethods of placing and holding face masks to minimisethe leak for Fisher & Paykel 60 mm and Laerdal size 0/1masks. 50 clinical staff gave positive pressure ventilationto a modified manikin designed to measure leak at theface mask. They were provided with written instructionson how to position and hold each mask and then receiveda demonstration. Face mask leak was measured aftereach teaching intervention.Results: A technique of positioning and holding the facemasks was identified which minimised leak. The mean(SD) mask leaks before instruction, after instruction andafter demonstration were 55% (31), 49% (30), 33% (26)for the Laerdal mask and 57% (25), 47% (28), 32% (30)for the Fisher & Paykel mask. There was no significantdifference in mask leak between the two masks. Writteninstruction alone reduced leak by 8.8% (CI 1.4% to 16.2%)for either mask; when combined with a demonstrationmask leak was reduced by 24.1% (CI 16.4% to 31.8%).Conclusion:Written instruction and demonstration of theidentified optimal technique resulted in significantlyreduced face mask leak.
In a study comparing the Laerdal and new Fisher &Paykel round face masks, we found that duringsimulated neonatal resuscitation mask leak variedwidely irrespective of the mask, operator experi-ence or technique of mask application.1 Theaverage mask leak was greater than 55% and thetechniques of mask application varied consider-ably.1
Successful neonatal face mask ventilationrequires an airtight seal between the rim of themask and the face.2–4 Achieving this can bedifficult.5 Efficacy of mask ventilation is judgedby observing chest rise and an increase in heartrate.6 Leak at the face mask is a common reason forfailure of ventilation2 and often occurs between thecheek and bridge of the nose at the orbitalmargin.3 4 It is a widespread misconception thatusing a manometer allows the operator to detectleaks.7 8 Even with large leaks, in the presence of ahigh gas flow into the system, the set pressure isachieved.1 8 9 Many participants in our previous
study were unaware of their leak and most wereunable to accurately assess the leak during positivepressure ventilation (PPV).1
Face mask ventilation is a mandatory skill forstaff caring for newborn infants10 and trainingemploys a variety of educational tools.2 3 11–14
Recommended techniques of mask placement andhold are variable,2–4 6 13 and have not been formallystudied.Round-shaped masks are the most widely used
designs,15–17 and are more easily cleaned.15 18 Acorrectly sized mask should cover the nose andmouth,2–4 6 19 without extending beyond the chintip or encroaching on the eyes.2–4 6 Applyingexcessive pressure may bruise the face and mouldthe back of the head.3 4
The primary aims of this study were to: (1)identify the technique for placing and holding aface mask which minimised mask leak duringsimulated neonatal resuscitation; (2) investigatethe effect of written instruction and a demonstra-tion of the identified technique on reducing maskleak from a Laerdal round mask and a Fisher &Paykel mask. A secondary aim was to determinewhether leak at the face mask differed between thetwo mask designs following training.
METHODSEquipmentFace masksThe size 0/1 Laerdal round mask (LR) (Laerdal,Stavanger, Norway) and the 60 mm Fisher &Paykel (FP) (Fisher & Paykel Healthcare,Auckland, New Zealand) ‘‘round’’ neonatal resus-citation mask were the appropriate sizes for themanikin. Participants had used the FP mask oncepreviously and the LR mask is in routine use at ourhospital.
Manual ventilation deviceThe Neopuff infant resuscitator (Fisher & PaykelHealthcare, Auckland, New Zealand) was used. Itis a continuous flow, pressure limited, T-piecedevice with a built-in manometer and a positiveend expiratory pressure (PEEP) valve. The operatorset the gas flow to 8 l/min, the peak inspiratorypressure (PIP) to 30 cm H2O and PEEP to 5 cmH2O.
Modified manikinA Laerdal Resusci baby manikin (Laerdal,Stavanger, Norway) was modified by replacingthe original ‘‘lung’’ with a 50 ml test lung (Drager,Lubeck, Germany) connected by non-distensible
Original article
F230 Arch Dis Child Fetal Neonatal Ed 2008;93:F230–F234. doi:10.1136/adc.2007.117788
A40
tubing to the mouth by an airtight seal. It was positioned soinflation caused chest rise. A pressure monitoring line wasconnected to the airway tubing immediately proximal to thetest lung. The compliance of this whole system whenpressurised to 30 cm H2O was 0.5 ml/cm H2O, with a maximaltest lung volume at this pressure of 65 ml.
Respiratory monitor, recording equipment and values measuredA Florian respiratory monitor (Acutronic Medical Systems, Ag,Switzerland) was used to measure inflating pressures and gasflow. The flow sensor was placed between the T-piece and facemask. The tidal volume passing through the sensor wascalculated by integrating the flow signals and calibrated witha 10 ml syringe. Percentage leak at the face mask was thedifference between the inspired and expired tidal volumes,expressed as a percentage of the inspired tidal volume(Percentage leak = [(inspiratory tidal volume – expiratory tidalvolume) 4 inspiratory tidal volume] 6100)Airway pressure was measured directly and calibrated against
a column of water.Mask leak rather than tidal volume is presented because it is
the best way to compare different techniques and masks withdifferent volumes and compliance.The Spectra software programme (Grove Medical, London,
UK) was used to acquire output from the Florian monitorthrough an analogue to digital converter onto a dedicatedcomputer. The Florian monitor and computer screen were notvisible to participants.
Determining the technique for mask placement and hold thatminimises leakThree experienced neonatologists investigated recommendedtechniques of positioning and holding face masks2–4 6 13 19 todetermine the method that minimised leak. These techniquesincluded: two ways of positioning a mask on the face; placing itstraight onto the face or rolling it onto the face beginning at thechin tip.3 4 There were three methods of holding the mask onthe face, which are shown in figure 1: ‘‘The stem hold’’—wherethe mask stem is held between the index finger and thumb.13
‘‘The two-point top hold’’—the thumb and index finger applybalanced pressure to the top flat portion of the mask where thesilicone is thickest. The stem is not held and the fingers shouldnot encroach onto the skirt of the mask.13 ‘‘The OK rim hold’’—the thumb and index finger form a C shape (as in the ‘‘OK’’hand gesture) placed around the top flat portion of the maskapplying an even distribution of pressure to the outer edge andnot encroaching onto the skirt of the mask.4 Jaw lift was alsoinvestigated; it was applied by using the third, fourth and fifthfingers to lift the chin forward4 against the downward pressureon the mask.
The three neonatologists tested the different components ofeach technique. They found that rolling the mask upwards ontothe face from the chin tip resulted in more accurate positioningof the mask and reduced the leak at the orbital margin. Jaw liftreduced the leak and also reduced the downward pressure ontothe manikin’s head. Rolling the mask onto the face and jaw liftwere subsequently used with each of the three mask holds.When they were satisfied they had perfected all techniques eachneonatologist ventilated the modified manikin using the threeholds with each mask five times (15 tests in total for each maskhold combination). Each recording lasted 1 minute. The resultsare shown in figure 2, where mean percentage mask leak isgrouped by type of hold for each mask design. The lowestmedian leak was found with the two-point top hold for the LRmask and with the rim hold for the FP mask (fig 3). During thisprocess two clinical cues, which identified the presence of agood mask seal, were identified; a whistling/hissing sound fromthe PEEP valve during expiration and the maintenance of PEEPat the set level.
Figure 1 Photographs of the three holdtypes demonstrated on the Laerdal roundneonatal mask.
Figure 2 Percentage leak at the face mask for each mask hold type andmask design (LR dark shading, FP light shading). Box plots show medianvalues (solid lines), interquartile range (margins of box), range of data,outliers (circles) and extreme values (asterisk).
Original article
Arch Dis Child Fetal Neonatal Ed 2008;93:F230–F234. doi:10.1136/adc.2007.117788 F231
A41
Setting and participantsThe 50 clinicians from our first study1 took part in this study.They included consultants, fellows, registrars, neonatal nursesand midwives of the Royal Women’s Hospital, Melbourne,Australia. Their years of neonatal experience were recorded.
Study protocolThe results of our first study1 provided the baseline percentagemask leak, ventilation rate, PIP and PEEP for both mask designsand for each participant using their usual techniques of maskapplication before any training; these data are presented in theresults as ‘‘before instruction.’’Each participant received written instructions containing text
and photographs of the techniques the investigators found tohave the least leak for each mask. Before each recording,participants were given time to familiarise themselves with theequipment. Participants were asked to ventilate the manikin for2 minutes at a rate of 40–60 inflations per minute with a PIP of30 cm H2O and a PEEP of 5 cm H2O ensuring that they wereachieving adequate chest rise. We measured the mean ventila-tion rate, PIP, PEEP, inspired and expired tidal volumes andpercentage leak for 2 minutes with each mask, excluding thefirst five inflations in each case.The best technique for each mask was then demonstrated to
each participant by one instructor (FEW). The inability toachieve the set PEEP with a large leak was demonstrated; as wasthe presence of a strong whistling/hissing noise from the PEEPvalve of the Neopuff, which occurs with minimal face maskleak. Recordings were then made for 30 seconds of PPV witheach mask, excluding the first three inflations in each case.Participants knew face mask leak was being assessed.
Participant surveyAt the end of the assessments participants were asked, ‘‘Wouldyou now feel more or less confident using this mask to deliverintermittent positive pressure ventilation (IPPV) in a resuscita-tion situation with the ‘two-point top hold’ for the LR maskand … with the ‘OK Rim hold’ for the FP mask.’’ There werefive possible responses ranging from much more confident tomuch less confident.
Randomisation and statistical analysisThe order of assessment of the two masks was determined by acomputer generated randomisation list. Data were analysedusing SPSS version 12.0 and Stata 9.2. Multiple linear regressionwas used to investigate the difference in face mask leak betweenthe two masks and to assess the influence of technique trainingon mask leak. The regression analysis uses robust standarderrors to allow for the repeated measures. Paired t tests wereused to compare other ventilation parameters. A p,0.05 was
considered significant. Data are expressed as mean (SD) orpercentage unless otherwise stated.
RESULTSParticipantsThere were 10 participants in each professional category. Themedian years of experience were: consultants 18.5 years, fellows(SpR level) 3.3 years, registrars (SHO level) 0.1 years, neonatalnurses 9.5 years and midwives 3.5 years. A total of 10 648inflations were recorded and analysed.The effect on mask leak for the three holds and two masks is
shown in figure 2. The median percentage leak for the two-point top hold and the rim hold was the smallest and leastvariable for the LR and FP masks, respectively. These two holdsare shown in figure 3. Participants’ usual techniques werepresented in our first study.1
Inflation pressures and rateThe mean (SD) PIP, PEEP and ventilation rates with the LRmask were 29.3 (2.8) cm H2O, 4.0 (0.7) cm H2O and 49 (19.8)/min before instruction and 29.7 (2.1) cm H2O, 4.0 (0.8) cm H2Oand 54 (36.9)/min after instruction. With the FP mask theywere 29.3 (1.8) cm H2O, 3.9 (0.7) cm H2O and 52 (38.1)/minbefore instruction and 29.7 (1.9) cm H2O, 4.1 (0.7) cm H2O and49 (13.7)/min after instruction. There was no significantdifference between the two masks before or after instruction.The only significant difference following instruction was for theFP mask where the PEEP increased by 0.2 cm H2O (p=0.038, CI0.1 to 0.4).
Percentage gas leak at the face maskThe mean (SD) percentage leak before instruction, afterinstruction and after demonstration of the optimal techniqueswas 55% (31), 49% (30), 33% (26) and 57% (25), 47% (28), 32%
Figure 3 Photographs of the optimal hold technique for each maskdesign.
Figure 4 Percentage leak at the face mask for each mask design whenused before instruction, after written instruction and after demonstrationof the techniques (LR dark shading, FP light shading). Box-plots showmedian values (solid line) interquartile range (margins of box), and rangeof data. ‘‘After demonstration’’ indicates the combined effect on maskleak of the previous written instruction and following demonstration.
Original article
F232 Arch Dis Child Fetal Neonatal Ed 2008;93:F230–F234. doi:10.1136/adc.2007.117788
A42
(30) for the LR and FP masks respectively. There was nosignificant difference between the two masks.Compared with the mean percentage mask leak of partici-
pants before instruction, with either mask, written instructionreduced mask leak by 8.8% (CI 1.4% to 16.2%, p=0.02), thedemonstration further reduced leak by 15.3% (CI 9.0% to22.7%, p,0.0005). The combined effect of written instructionfollowed by a demonstration reduced leak by 24.1% (CI 16.4%to 31.8%, p,0.0005). Figure 4 shows a box-plot of percentagemask leak for each mask design before and after training.
Participant survey responsesFor the LR mask using the ‘‘two-point top hold,’’ 32% weremuch more/more confident, 50% neutral/no change and 18%much less/less confident. For the FP mask with ‘‘the OK rimhold,’’ 56% were much more/more confident, 24% neutral/nochange and 20% much less/less confident.
DISCUSSIONThis study identified techniques of placing and holding neonatalface masks which minimised mask leak. Teaching participantsthese techniques significantly reduced the leak when theyventilated a modified manikin. There was no significantdifference in face mask leak between the two neonatal masksbefore or after technique training.With variable and unmeasured mask leaks variable tidal
volumes are delivered to the lungs. A large leak may reduce thetidal volume and a small leak, with a fixed inflating pressure,may result in excessive tidal volumes. Successful neonatalresuscitation is dependent on delivering an appropriate tidalvolume. Controlling the face mask leak is one part of achievingthis.This study was done using a modified manikin; the results
may be different with newborn infants.A criticism of this study is that the reduction in leak may
have resulted from the experience gained by the participants asthe studies progressed. If this were the case it is likely that themost experienced clinicians would have lower leaks; howeverwe found a wide and similar variation across all professionalcategories. This variation in performance irrespective ofexperience was previously reported.9 20 If experience wasimportant then we would have expected the leak to be lowerwith a familiar mask than with a new mask; this was not thecase.Caution should be applied in extrapolating the results of this
study. Different devices have different characteristics—forexample, self inflating bags without PEEP will have a lowerleak. The use of different inflation times (Ti) will also alter leak.We did not measure Ti in our study but it is unlikely that thissystematically changed through the different phases of thestudy.Clinical skills are best taught in stages and both theoretical
knowledge and practical demonstration are important.3 13 21 Wehave seen that a significant reduction in leak can be achievedwith written instruction alone; with a further importantreduction in leak after the technique was demonstrated.It has been shown that the manometer does not reliably
identify mask leak unless it is very large.1 8 9 During face maskventilation, poor mask application and resultant variable maskleaks affect the tidal volume delivery, which may be inadequatefor effective gas exchange.7 22 The delivery of very low tidalvolumes as a result of large mask leaks could result in failure to
resuscitate an infant. Clinical cues, which suggest mask leak,were identified and used in the education process.This study has shown that identifying, teaching and
demonstrating improved techniques of placing and holdingneonatal face masks reduces leak with a modified manikin. Thismay lead to improvements in the efficacy of face maskventilation during neonatal resuscitation and hopefully areduction in the incidence of failed face mask resuscitation.
CONCLUSIONThere was no significant difference in face mask leak betweenthe Laerdal and Fisher & Paykel neonatal face masks whentested on a modified manikin. Specific techniques of face maskplacement and hold reduced leak and were incorporated intoeducational strategies. There was a significant reduction in facemask leak following written instruction alone and a furtherreduction following demonstration. Further studies are requiredto determine whether these techniques prove effective in thedelivery room.
Acknowledgements: We thank the staff of the Royal Women’s Hospital Melbournefor their participation, and Fisher & Paykel Healthcare for providing samples of theirneonatal resuscitation mask. Supported in part by a RWH Postgraduate degreescholarship (COFK) and a NHMRC Practitioner Fellowship (PGD); the AustralianNational Health and Medical Research Council Program Grant No 384100.
Funding: None.
Competing interests: None.
REFERENCES1. Wood FE, Morley CJ, Dawson JA, et al. Assessing the effectiveness of two round
neonatal resuscitation masks: study 1. Arch Dis Child Fetal Neonatal Ed2008;93:F235–7.
2. Resuscitation Council (UK). Resuscitation at birth: the newborn life supportprovider manual. London: Resuscitation Council (UK), 2001.
3. American Academy of Pediatrics/American Heart Association. Textbook ofneonatal resuscitation, 4th ed. Elk Grove Village, IL: American Academy of Pediatrics/American Heart Association, 2000.
4. American Academy of Pediatrics/American Heart Association. Use ofresuscitation devices for positive-pressure ventilation. In: Kattwinkel J, ed. Textbookof neonatal resuscitation. Elk Grove Village, IL: American Academy of Pediatrics/American Heart Association, 2006.
5. Finer NN, Rich W, Craft A, et al. Comparison of methods of bag and mask ventilationfor neonatal resuscitation. Resuscitation 2001;49:299–305.
6. Kattwinkel J, Niermeyer S, Nadkarni V, et al. An advisory statement from thePediatric Working Group of the International Liaison Committee on Resuscitation.Pediatrics 1999;103:e56–8.
7. O’Donnell CP, Davis PG, Morley CJ. Resuscitation of premature infants: what are wedoing wrong and can we do better? Biol Neonate 2003;84:76–82.
What is already known on this topic
c Leak at the face mask is large during PPV and may result ininadequate ventilation.
c Operators are unable to reliably detect leak during PPV.c Techniques used to position and hold masks on the face are
variable.
What this study adds
c No difference in mask leak was demonstrated between tworound face masks before or after technique training.
c Teaching an improved technique significantly reduces leak atthe face mask when tested with a modified manikin.
Original article
Arch Dis Child Fetal Neonatal Ed 2008;93:F230–F234. doi:10.1136/adc.2007.117788 F233
A43
8. O’Donnell CP, Davis PG, Lau R, et al. Neonatal resuscitation 3:manometer use in a model of face mask ventilation. Arch Dis Child Fetal Neonatal Ed2005;90:F397–400.
9. O’Donnell CP, Davis PG, Lau R, et al. Neonatal resuscitation 2: an evaluation ofmanual ventilation devices and face masks. Arch Dis Child Fetal Neonatal Ed2005;90:F392–6.
10. O’Donnell CP, Stewart MJ, Mildenhall LF. Neonatal resuscitation in Australia andNew Zealand. J Paediatr Child Health 2006;42:4–5.
11. Baskett PJ, Nolan JP, Handley A, et al. European Resuscitation Council guidelinesfor resuscitation 2005. Section 9. Principles of training in resuscitation. Resuscitation2005;67(Suppl 1):S181–9.
12. Jewkes F, Phillips B. Resuscitation training of paediatricians. Arch Dis Child2003;88:118–21.
13. Members of the Newborn Life Support Working Party.The newborn life support instructor manual: generic instructor course.Resuscitation Council UK, 2001.
14. Australian Resuscitation Council. Neonatal guidelines. ARC manual ofguidelines.Melbourne, Australia: Australian Resuscitation Council, 2006.
15. O’Donnell CP, Davis PG, Morley CJ. Neonatal resuscitation: review of ventilationequipment and survey of practice in Australia and New Zealand. J Paediatr ChildHealth 2004;40:208–12.
16. O’Donnell CP, Davis PG, Morley CJ. Positive pressure ventilation at neonatalresuscitation: review of equipment and international survey of practice. Acta Paediatr2004;93:583–8.
17. Leone TA, Rich W, Finer NN. A survey of delivery room resuscitation practices in theUnited States. Pediatrics 2006;117:e164–75.
18. Palme C, Nystrom B, Tunell R. An evaluation of the efficiency of face masks in theresuscitation of newborn infants. Lancet 1985;1:207–10.
19. Field D, Milner AD, Hopkin IE. Efficiency of manual resuscitators at birth. Arch DisChild 1986;61:300–2.
20. Redfern D, Rassam S, Stacey MR, et al. Comparison of face masks in the bag-maskventilation of a manikin. Eur J Anaesthesiol 2006;23:169–72.
21. Goldmann K, Ferson DZ. Education and training in airway management. Best PractRes Clin Anaesthesiol 2005;19:717–32.
22. Milner AD, Vyas H, Hopkin IE. Efficacy of facemask resuscitation at birth. BrMed J (Clin Res Ed) 1984;289:1563–5.
BMJ Careers online re-launches
BMJ Careers online has re-launched to give you an even better online experience. You’ll still find ouronline services such as jobs, courses and careers advice, but now they’re even easier to navigate andquicker to find.
New features include:c Job alerts – you tell us how often you want to hear from us with either daily or weekly alertsc Refined keyword searching making it easier to find exactly what you wantc Contextual display – when you search for articles or courses we’ll automatically display job adverts
relevant to your searchc Recruiter logos linked directly to their organisation homepage – find out more about the company
before you applyc RSS feeds now even easier to set up
Visit careers.bmj.com to find out more.
Original article
F234 Arch Dis Child Fetal Neonatal Ed 2008;93:F230–F234. doi:10.1136/adc.2007.117788
A44
Ventilation and Spontaneous Breathing at Birth of Infants with CongenitalDiaphragmatic Hernia
ARJAN B. TE PAS, MD, C. OMAR F. KAMLIN, JENNIFER A. DAWSON, MN, COLM O’DONNELL, PHD, JENNIFER SOKOL, MICHAEL STEWART,COLIN J. MORLEY, MD, AND PETER G. DAVIS, MD
Objective To describe the interaction of spontaneous breaths, manual ventilation, and tidal volumes (VT) during stabili-zation of infants with congenital diaphragmatic hernia (CDH) in the delivery room.Study design We studied infants with CDH receiving respiratory support at birth. Airway pressure, flow, and volume weremeasured, and each breath or inflation was analyzed. Each VT was classified as a manual inflation, a spontaneous breath, ora spontaneous breath coinciding with manual inflation on the basis of the timing of the pressure and flow waves.Results Twelve infants had 2957 breaths suitable for analysis, with spontaneous breathing in 11 infants (92%). The mean(!SD) proportion of manual inflations was 41% (!24%), spontaneous breaths 43% (!25%), spontaneous but coinciding withmanual inflation 16% (!12%). VT was significantly different for spontaneous breaths (3.8 ! 1.9 mL/kg), spontaneous breathscoinciding with manual inflation (4.7 ! 2.5 mL/kg), and manual inflations alone (2.6 ! 1.6 mL/kg).Conclusions Most infants with CDH breathed spontaneously, and manual ventilation was mostly asynchronous. Weobserved large differences in tidal volumes between spontaneous breaths, manual inflations, or where these coincided, withmanual inflations having the lowest VT. Monitoring the respiratory pattern of these infants could improve respiratory support.(J Pediatr 2009;154:369-73)
Congenital diaphragmatic hernia (CDH) occurs in approximately 1 per 2500 live births.1 Mortality and morbidity ratesfrom CDH are substantial and are related to the degree of pulmonary hypoplasia,1 size of diaphragmatic defect,2,3 andassociated anomalies.4 Observational studies have demonstrated improvements in patient survival rates with ventilation
strategies, which might limit overdistension and barotrauma, that encourage spontaneous breathing.5-8 However, no studies havereported detailed measurements of respiratory variables immediately after delivery. Stabilization of infants with CDH in thedelivery room is guided by protocols and commentary from experts.9,10 The InternationalLiaison Committee on Resuscitation11 has not made any specific recommendations on themanagement of infants with CDH in the delivery room.
Guidelines that suggest upper limits of peak inspiratory pressure (PIP) may beinappropriate because lung compliance varies considerably between individual infants andin the same infant in the first minutes of life. Changing the PIP to target a tidal volume(VT) may be more logical, but there are no data indicating the range of tidal volumes foran infant with CDH. In addition, an infant with CDH often breathes at birth and willinfluence the amount of respiratory assistance required. The aim of this study was todescribe the interaction of spontaneous breaths and manual ventilation and to measuretidal volumes during stabilization of infants with CDH immediately after birth.
METHODSThe Royal Women’s Hospital (RWH) Melbourne is a tertiary level perinatal center
with approximately 6500 deliveries per year and is a regional referral center for antenatalmanagement of CDH. All live births of infants known to have CDH between February2004 and August 2007 were eligible for inclusion if a member of the research team wasavailable to attend the delivery. After birth infants were admitted to the neonatal
CDH Congenital diaphragmatic herniaCPAP Continuous positive airway pressurePEEP Positive end-expiratory pressure
PIP Peak inspiratory pressureRWH Royal Women’s HospitalVT Tidal volume
From the Division of Newborn Services,Royal Women’s Hospital (A.T., C.K., J.D.,J.S., M.S., C.M., P.D.), Victoria, Australia, theDivision of Neonatology, Department ofPediatrics, Leiden University Medical Cen-ter (A.T.), Leiden, The Netherlands, andthe Department of Neonatology, NationalMaternity Hospital (C.O.), Dublin, Ireland.C.O.F.K., J.D., and A.t.P., have been benefi-ciaries of the Royal Women’s HospitalPostgraduate Degree Scholarship. P.G.D., ispartly supported by a National Health andMedical Research Council Fellowship. Theauthors declare no conflicts of interest.Submitted for publication May 15, 2008;last revision received Jul 25, 2008; acceptedSep 12, 2008.Reprint requests: Arjan B. te Pas, MD, Di-vision of Neonatology, Department of Pe-diatrics, Leiden University Medical Center,J6-S, P.O. Box 9600, 2300 RC Leiden, TheNetherlands. E-mail: [email protected]/$ - see front matterCopyright © 2009 Mosby Inc. All rightsreserved.10.1016/j.jpeds.2008.09.029
369
A45
intensive care unit and stabilized before being transferredto the regional surgical center (Royal Children’s Hospital,Melbourne).
Resuscitation PracticeAll staff complete an in-house resuscitation training
program based on the Australasian Resuscitation Councilguidelines.12 The resuscitation team attending the birth of aninfant with CDH included a consultant neonatologist, a neo-natal fellow, a resident, a neonatal nurse, and a member of theResuscitation Research team.
For manual ventilation we used the Neopuff InfantResuscitator (Fisher & Paykel, Auckland, New Zealand).This is a manually operated device that delivers gas flowthrough a T-piece in a mask or endotracheal tube. TheT-piece has a valve that controls the positive end-expiratorypressure (PEEP). PIP is generated by occlusion of the valvewith a finger. PIP and PEEP are set by the operator and forrespiratory support of infants with CDH set at 25 and 5 cmH2O, respectively.
The delivery room management of infants with CDHwas individualized, according to the infants’ condition andresponse to initial resuscitation and based on expert opinion.9
When the infant was breathing spontaneously without sig-nificant respiratory distress, continuous positive airway pres-sure (CPAP) via mask was given. When there were signs ofrespiratory difficulty, infants were intubated orally and man-ually ventilated at the earliest opportunity. Endotracheal tubeposition was judged from clinical signs and an exhaled carbondioxide detector (Pedi-Cap, Nellcor Puritan Bennett, Pleas-anton, California). Positive pressure ventilation by mask wasavoided if possible. All infants were electively intubatedwithin 10 minutes of age according to protocol. Ventilationwas provided with the Neopuff until the infant was trans-ferred to the transport ventilator.
Preductal oximetry with the Masimo Radical (MasimoCorp, Irvine, California) was obtained with a sensor on theinfant’s right hand. The oximeter was set to capture heart rateand oxygen saturation with maximal sensitivity and averagedata over 2-second intervals. Resuscitation was started ini-tially with 100% of oxygen until the RWH policy waschanged in 2006 in line with the Australian Neonatal Resus-citation Guidelines.12 There is accumulating evidence thatresuscitation with air is as effective as starting with 100%oxygen and that resuscitation of infants with 100% oxygenincreases the mortality rate and delays the onset of breath-ing.13 A preductal saturation range of 75% to 85% was tar-geted. A large-bore nasogastric tube was passed to decom-press the stomach. Central arterial and venous access wasdeferred (unless clinically indicated) until admission to theneonatal intensive care unit and a blood gas reading wasperformed.
Recording Respiratory VariablesInspiratory and expiratory gas flow, VT, and pressure
were measured with the Florian Respiratory Monitor (Acu-
tronic, Zug, Switzerland). This measures pressure in therespiratory circuit directly (at the gas outlet) and uses a sensorplaced between the ventilation device and the endotrachealtube to measure gas flow. The data were recorded and ana-lyzed with Spectra software (Grove Medical, London, UnitedKingdom), a program designed specifically for the recordingand analysis of neonatal respiratory signals. The resuscitationteam was not blinded to the Florian display, but the team wasnot informed of the measured tidal volumes and made theirdecisions on clinical assessment (heart rate, chest movementand oxygen saturations).
A priori, we decided to limit data collection to 10minutes after cord clamping. The variation in the amount ofdata collected were dependent on the duration of ventilation,intubation attempts, confirmation of the endotracheal tubeposition and the time taken to resite the flow sensor afterinterventions. Breath-by-breath analysis was performed man-ually to identify one of 3 respiratory patterns:
MANUAL INFLATION. This was characterized by an inspiratoryflow concurrent with an obvious increase in airway pressurefrom an inflation, the start of expiratory flow synchronouswith end of the inflation and at the same time the airwaypressure starts to return to the baseline pressure (Figure 1).
SPONTANEOUS BREATH. This was characterized by an inspira-tory and expiratory flow in absence of a concurrent pressure
Figure 1. Two examples of manual inflations, characterized by aninspiratory flow concurrent with an obvious increase in airway pressurefrom an inflation, the start of expiratory flow synchronous with end of theinflation and at the same time the airway pressure starts to return to thebaseline pressure. In the first example (patient 1), expiratory flow does notreturn to zero which indicates some air-trapping, however this is notconfirmed by the expiratory volume. The second example is of patient 9,despite use of high pressures very little flow and volume are reached. Thisinfant died before transfer to the surgical unit.
370 te Pas et al The Journal of Pediatrics • March 2009
A46
wave from an inflation. Spontaneous breaths occurred during(endotracheal) CPAP and in between inflations (Figure 2).
SPONTANEOUS BREATH COINCIDING WITH A MANUAL INFLATION.This was characterized by a combined pattern with juxtaposedor synchronized spontaneous breath during a manual infla-tion, characterized by the occurrence of 2 coincidental in-spiratory flow patterns resulting in 1 volume wave (Figure 3).
For each pattern, the Spectra program was used to mea-sure expired tidal volumes and minute ventilation. Breaths andinflations were excluded from final analysis if leakage was greaterthan 20% ([VT inspired ! VT expired]/VT inspired " 100).
Statistical AnalysisBecause no information was available on which to base
a sample size calculation, we recruited a convenience sampleover a 3.5-year period. Parametric and nonparametric testswere used for normally distributed and skewed data respec-tively. Tidal volumes of the 3 different types of breaths werecompared by use of 1-way analysis of variance and post-hoccomparison was performed by use of the Tukey HSD test.Data were analyzed with SPSS version 15 (SPSS Inc, Chi-cago, Illinois).
EthicsThis study was part of a prospective observational study
with recordings of delivery room resuscitations with the en-dorsement of the Human Research and Ethics Committee at
RWH. Antenatal consent was obtained before birth if themother was not in established labor and if time permitted.Where this was not possible, a waiver of consent was usedaccording to the Australian NHMRC guidelines for studiesin emergency medicine, and retrospective consent was soughtfrom the parents to use the recording.
RESULTSOver a 3.5-year period, 31 infants with antenatally
diagnosed CDH were born alive at RWH. The research teamwas able to attend 22 deliveries. At 5 deliveries there wasinsufficient time to set up the equipment for physiologicalmonitoring, and 5 recordings were excluded for poor quality.The remaining 12 infants formed the study population.
The demographic characteristics of the study groupwere as follow (expressed as mean [#SD] or median [range]where appropriate): gestational age 39 (#1.0) weeks, birthweight 3.3 (#0.3) kg, 5-minute Apgar score 9 (2-10), cae-sarean section 60%, left-sided lesion 90%, pulse oximetry at 5minutes and 10 minutes, respectively, 65% (#29%) and 84%(#23%), heart rate at 5 and 10 minutes, respectively, 138(#49) and 163 (#29) beats/min, at admission pH 7.19(#0.23), partial pressure of carbon dioxide in arterial blood of53 (34-161) mm Hg, and fractional concentration of oxygenin inspired gas of 23% (21%-100%). Ventilation settings in
Figure 2. Two examples of spontaneous breaths of patient 10. In the firstexample the infant is breathing on CPAP of 8 cm H2O, inspiratory flowis concurrent with a decrease in the CPAP-pressure and expiratory flowincreases the CPAP-pressure. In the second example the infant isventilated with a Laerdal bag and mask. Spontaneous breaths are taken inbetween the inflation and have a larger tidal volume than when inflation isgiven.
Figure 3. Three examples of spontaneous breath coinciding with amanual inflation, juxtaposed or synchronized with a mandatory inflation,characterized by the occurrence of two peak inspiratory flow patternsresulting in one volume wave. In the first 2 examples (patient 8 and 11)inspiratory flow start before inflation is given, when inflation start a secondpeak in inspiratory flow is visible. In the third example (patient 2)spontaneous breath occurs right in the middle of the inflation observed asa second peak in inspiratory flow and decrease in pressure, but resulting toone volume wave.
Ventilation and Spontaneous Breathing at Birth of Infants with Congenital Diaphragmatic Hernia 371
A47
the neonatal intensive care unit were as follows: volumeguarantee 3.8 (#0.5) mL/kg, peak inspiratory pressures mea-sured 29 (#9) cm H2O and maximum set ventilator rates 50(#8) min!1. Two infants had an associated anomaly: 1 withtalipes equinovarus and another with postnatally diagnosedcomplex congenital heart disease (Table). Karyotype was nor-mal in all infants tested. No infants received antenatal ste-roids. Face mask continuous positive airway pressure (CPAP)was provided as initial respiratory support in 4 infants withthe remainder electively intubated. In between endotrachealintubation attempts, 3 infants received intermittent positivepressure ventilation with the Neopuff Infant Resuscitator andfacemask. There was 1 death before transfer to the surgicalunit, and 1 infant died after withdrawal of care. Ten infantssurvived to hospital discharge after surgery. Three infantsrequired home oxygen therapy after discharge (Table).
Breathing PatternsFrom the 12 infants, 3547 breaths or inflations were
recorded. Leak in excess of 20% was seen in 603 (17%) andthese were excluded. A total of 2957 breaths or inflations wereanalyzed, with an average of 237 (85) breaths per patient(Table).
Spontaneous breathing was observed in 11 infants(92%). Overall mean (#SD) proportion of manual inflationswas 41% (#25%). Spontaneous breaths accounted for 43%(#25%) and spontaneous breaths that coincided with aninflation accounted for 16% (#12%). The percentages ofthese patterns in each patient are shown in (Table).
Respiratory VariablesIn all breaths and inflations of the 12 patients the mean
(#SD) respiratory rate was 65 (#24) breaths/min, tidal vol-ume 3.4 (#2.1) mL/kg, minute volume 210 (#129) mL/kg/min, inspiration time 0.41 (#0.20) s, expiration time 0.60
(#0.33) s, peak inspiratory flow 4.1 (2.1) L/min and peakexpiratory flow 3.7 (1.8) L/min. The mean (SD) peak inspira-tory pressure was 26.1 (2.1) cm H2O and PEEP was 5.5 (1.7)cm H2O.
Tidal VolumeThe mean (#SD) VT during spontaneous breaths was 3.8
(#1.9) mL/kg, during spontaneous breaths that coincided withinflation it was 4.7 (#2.5) mL/kg and during manual inflationsit was 2.6 (#1.6) mL/kg. A 1-way analysis of variance showed astatistically significant difference in tidal volume for the 3 groups(F[2, 2873] $ 281.2; P % .0001). Post-hoc comparison withTukey HSD test indicated that all groups were significantlydifferent from each other (P % .0001). The mean VT for eachpattern per patient are shown in the Table.
DISCUSSIONWe reported respiratory patterns and variables during
stabilization of infants with CDH at birth. Most infants inour cohort breathed at birth, and only a small proportion ofthese spontaneous breaths coincided with a manual inflation.Significantly higher VT occurred during spontaneous breathswith or without coinciding inflations, compared with manualinflations alone.
Current guidelines for the management of infants withCDH advise limiting the peak inspiratory pressure (usuallynot exceeding 25 cm H2O) and advocate permissive hyper-capnia to avoid overdistention.14,15 However, there is a grow-ing appreciation that in preterm infants volutrauma may bemore important than barotrauma,16-18 and this may also bepertinent to the management of infants with CDH. The VTdelivered when a set pressure is used depends on the compli-ance of the lung and the spontaneous respiratory effort of theinfant. The lung compliance of an infant with CDH is mainlydetermined by the severity of the lung hypoplasia. In addition,
Table. Patterns, expired tidal volume, and minute volume per patient
PatientBW(kg)
Breathsanalyzed
(N)
Mandatoryinflations (%),
Tve (SD)mL/kg
Spontaneousbreaths (%),
Tve (SD)mL/kg
Breathscoinciding
inflation (%),Tve (SD)
mL/kgMV (SD)
mL/kg Anomalies
Survival tohospital
dischargeHomeoxygen
1 3.2 382 53%, 3.2 (1.5) 19%, 4.1 (1.5) 28%, 4.8 (2.0) 236 (95) Yes2 3.2 296 47%, 3.1 (1.5) 22%, 3.9 (2.3) 31%, 4.2 (1.6) 204 (100) Yes3 3.9 290 23%, 2.7 (1.2) 69%, 4.0 (1.0) 8%, 4.3 (1.0) 187 (85) Talipes Yes4 3.2 322 15%, 2.6 (0.6) 78%, 4.5 (1.5) 7%, 3.9 (1.3) 302 (106) Yes5 3.6 197 4%, 4.5 (2.3) 77%, 4.8 (1.5) 19%, 4.1 (2.2) 268 (117) Yes Home oxygen6 2.7 117 60%, 3.7 (1.1) 40%, 3.6 (1.7) 0% 207 (82) Yes Home oxygen7 2.8 198 31%, 1.9 (0.8) 68%, 1.8 (0.9) 1%, 2.0 (0.8) 114 (57) Yes8 3.0 285 50%, 2.2 (0.8) 24%, 2.9 (1.2) 26%, 3.7 (1.1) 176 (74) CHD No9 3.3 113 100%, 0.6 (0.3) 0% 0% 70 (20) No
10 3.8 154 41%, 2.5 (1.0) 34%, 2.5 (0.6) 25%, 3.9 (0.6) 206 (61) Yes Home oxygen11 3.6 213 38%, 4.2 (2.7) 36%, 5.7 (3.5) 26%, 9.5 (0.4) 383 (197) Yes12 3.4 286 31%, 1.8 (0.8) 50%, 2.6 (1.2) 19%, 2.7 (1.7) 147 (76) Yes
Tve, Expired tidal volume; MV, minute volume.
372 te Pas et al The Journal of Pediatrics • March 2009
A48
the mechanical properties of the lung can change quickly asfluid is reabsorbed and the lung aerates. This makes it difficultto generalize and advise a “safe” range of VT for all infantswith CDH. Because volumes of both ipsilateral and contralat-eral lungs of infants with CDH appear to be lower thannormal,19 this makes it likely that the “safe” range of VTduring ventilation is lower than that for normal term infants(4-6 mL/kg).20 The individual VT measured during sponta-neous breathing may represent the “safe” range for a giveninfant. From our observations the implication is that usingpressures, as advised by protocol, in most cases the VT givenwas too low during a manual inflation and too high during asynchronized breath. Our speculation is that this could lead tovolutrauma.
Infants with CDH can be severely compromised, but inmost cases their respiratory effort is comparable to healthynewborn infants. Most infants started breathing at birth, andmost of the time manual inflations were not coincidental withthe infants’ breaths. This is comparable to nonsynchronizedventilation, and this may lead to inadvertent high pressuresand increases the risk for injury, such as pneumothoraxes.21 Itcould also lead to the infant “fighting against the ventilator,”which could aggravate the pulmonary hypertension. Althoughsome centers avoid these problems by sedating and musclerelaxing the infant,8,10 preservation of breathing has becomepart of the “gentle ventilation” protocols.5,6,22 We speculatethat the use of triggered (ie, synchronized) ventilation at birthin infants with CDH at birth could lead to more efficientsupport and reduces the risk of air leak.
The study is limited by the small number of infantsobserved. Therefore the results may not be representative ofall infants with CDH. However, we have shown that itfeasible to monitor respiratory function in infants with CDHat birth. Most clinicians in CDH centers are familiar withrespiratory function monitoring, and larger observationalstudies should be performed to better understand the require-ments for respiratory support.
REFERENCES1. Langham MR Jr, Kays DW, Ledbetter DJ, Frentzen B, Sanford LL, Richards DS.Congenital diaphragmatic hernia. Epidemiology and outcome. Clin Perinatol 1996;23:671-88.
2. Rygl M, Pycha K, Stranak Z, Melichar J, Krofta L, Tomasek L, et al. Congenitaldiaphragmatic hernia: onset of respiratory distress and size of the defect: analysis of theoutcome in 104 neonates. Pediatr Surg Int 2007;23:27-31.3. Lally KP, Lally PA, Lasky RE, Tibboel D, Jaksic T, Wilson JM, et al. Defect sizedetermines survival in infants with congenital diaphragmatic hernia. Pediatrics 2007;120:e651-7.4. Borys D, Taxy JB. Congenital diaphragmatic hernia and chromosomal anomalies:autopsy study. Pediatr Dev Pathol 2004;7:35-8.5. Boloker J, Bateman DA, Wung JT, Stolar CJ. Congenital diaphragmatic hernia in120 infants treated consecutively with permissive hypercapnea/spontaneous respiration/elective repair. J Pediatr Surg 2002;37:357-66.6. Wilson JM, Lund DP, Lillehei CW, Vacanti JP. Congenital diaphragmatichernia—a tale of two cities: the Boston experience. J Pediatr Surg 1997;32:401-5.7. Wung JT, Sahni R, Moffitt ST, Lipsitz E, Stolar CJ. Congenital diaphragmatichernia: survival treated with very delayed surgery, spontaneous respiration, and no chesttube. J Pediatr Surg 1995;30:406-9.8. Bagolan P, Casaccia G, Crescenzi F, Nahom A, Trucchi A, Giorlandino C.Impact of a current treatment protocol on outcome of high-risk congenital diaphrag-matic hernia. J Pediatr Surg 2004;39:313-8.9. Bohn D. Congenital diaphragmatic hernia. Am J Respir Crit Care Med 2002;166:911-5.10. Finer NN, Tierney A, Etches PC, Peliowski A, Ainsworth W. Congenitaldiaphragmatic hernia: developing a protocolized approach. J Pediatr Surg 1998;33:1331-7.11. The International Liaison Committee on Resuscitation (ILCOR) consensus onscience with treatment recommendations for pediatric and neonatal patients: neonatalresuscitation. Pediatrics 2006;117:e978-e988.12. Morley C. New Australian Neonatal Resuscitation Guidelines. J Paediatr ChildHealth 2007;43:6-8.13. Davis PG, Tan A, O’Donnell CP, Schulze A. Resuscitation of newborn infantswith 100% oxygen or air: a systematic review and meta-analysis. Lancet 2004;364:1329-33.14. Logan JW, Rice HE, Goldberg RN, Cotten CM. Congenital diaphragmatichernia: a systematic review and summary of best-evidence practice strategies. J Perinatol2007;27:535-49.15. Logan JW, Cotten CM, Goldberg RN, Clark RH. Mechanical ventilation strat-egies in the management of congenital diaphragmatic hernia. Semin Pediatr Surg2007;16:115-25.16. McCallion N, Lau R, Dargaville P, Morley CJ. Volume guarantee ventilation,interrupted expiration and expiratory braking. Arch Dis Child 2005;90:865-70.17. Dreyfuss D, Soler P, Basset G, Saumon G. High inflation pressure pulmonaryedema. Respective effects of high airway pressure, high tidal volume, and positiveend-expiratory pressure. Am Rev Respir Dis 1988;137:1159-64.18. Hernandez LA, Peevy KJ, Moise AA, Parker JC. Chest wall restriction limits highairway pressure-induced lung injury in young rabbits. J Appl Physiol 1989;66:2364-8.19. Peralta CF, Jani J, Cos T, Nicolaides KH, Deprest J. Left and right lung volumesin fetuses with diaphragmatic hernia. Ultrasound Obstet Gynecol 2006;27:551-4.20. Long EC, Hull WE. Respiratory volume-flow in the crying newborn infant.Pediatrics 1961;27:373-7.21. Greenough A, Dimitriou G, Prendergast M, Milner AD. Synchronized mechan-ical ventilation for respiratory support in newborn infants. Cochrane Database Syst Rev2008;CD000456.22. Sakurai Y, Azarow K, Cutz E, Messineo A, Pearl R, Bohn D. Pulmonarybarotrauma in congenital diaphragmatic hernia: a clinicopathological correlation. J Pe-diatr Surg 1999;34:1813-7.
Ventilation and Spontaneous Breathing at Birth of Infants with Congenital Diaphragmatic Hernia 373
A49
Spontaneous Breathing Patterns of Very Preterm Infants TreatedWith Continuous Positive Airway Pressure at Birth
ARJAN B. TE PAS, PETER G. DAVIS, C. OMAR F. KAMLIN, JENNIFER DAWSON, COLM P. F. O’DONNELL,AND COLIN J. MORLEY
Division of Newborn Services [A.B.P., P.G.D., C.O.F.K., J.D., C.J.M.], Royal Women’s Hospital, Carlton, Victoria 3053, Australia;Department of Neonatology [C.P.F.O.], National Maternity Hospital, Dublin 2, Ireland
ABSTRACT: There are no data describing how very preterm infantsbreathe spontaneously immediately after birth. We studied a conve-nience sample of spontaneously breathing infants !32 wk’ gestationtreated with facemask continuous positive airway pressure at birth.Airway pressure and flow were measured and each breath analyzed.Twelve infants had 792 breaths suitable for analysis. Results aregiven as mean (SD). Gestational age and birth weight were 29 (1.9)wk and 1220 (412) g. Recordings were started 159 (77) s after birth.The inspiratory pattern and duration was similar in all breaths at 0.36(0.11) s. There were five expiratory patterns; most infants had morethan one. In 79% of breaths expiratory duration (1.6 (1.1) s) wasslowed or held by interruption or braking of expiratory flow. It wasbraked in 47% to a complete expiratory hold, in 22% by grunting orcrying, and in 10% by slow or interrupted expiration. In 21% of thebreaths, expiration was not interrupted and lasted 0.53 (0.13) s. Half ofthese breaths represented a panting pattern (rate !60 /min). Imme-diately after birth, most very preterm infants, treated with continuouspositive airway pressure, frequently prolong their expiration bybraking the expiratory flow. (Pediatr Res 64: 281–285, 2008)
Very preterm infants may have difficulty aerating theirlungs after birth because of poor respiratory drive, weak
muscles, flexible ribs, surfactant deficiency, and impaired lungliquid clearance (1–4). However, many very preterm infantstreated immediately after birth with continuous positive air-way pressure (CPAP) breathe well and achieve normal bloodgases (5,6).
Between 1960 and 1986, observational data were gatheredimmediately after birth from small numbers of spontaneouslybreathing term infants and used to inform international guide-lines for neonatal resuscitation (7–14). There are no datadescribing how very preterm infants breathe and aerate theirlungs immediately after birth. It may be inappropriate toassume the results of studies of term infants also apply to verypreterm infants.
The aim of this study was to investigate the spontaneousbreathing patterns of very preterm infants treated with CPAPimmediately after birth.
PATIENTS AND METHODS
These physiologic studies 1were approved by the Royal Women’s HospitalResearch and Ethics committees. Parental consent for these recordings wasobtained.
At the Royal Women’s Hospital, Melbourne, spontaneously breathingnewly born very preterm infants with respiratory difficulty are assisted with afacemask CPAP. Between 2004 and 2007, we recorded physiologic parame-ters during delivery room stabilization of very preterm infants. All recordingsof infants born at !32 wk’ gestation, who were spontaneously breathing,while treated with CPAP via a face mask, were identified and their breathingpatterns in the first 10 minutes of life were analyzed. Our guidelines recom-mend to start with a CPAP-level set on 8 cmH2O. Infants were excluded ifpositive pressure ventilation was given at birth. We only used recordings ofinfants if there were: a) no signs of mask leak, b) a clean flow signal, notdisturbed by secretions or infant movements, and c) no signs of movement ofthe mask.
Recording equipment. Gas flow in and out of the infant was measuredwith a hot-wire anemometer (Florian: Acutronic Medical Systems AG, Zug,Switzerland) placed between the T-piece of a resuscitation device (NeopuffInfant Resuscitator, Fisher and Paykel, Auckland, New Zealand) and thefacemask. This signal was integrated to derive inspired and expired tidalvolume. Airway pressure was measured immediately proximal to the mask.The signals of airway flow, tidal volumes, and airway pressure were digitisedand recorded at 200 Hz using a neonatal respiratory physiologic recordingprogram (Spectra, Grove Medical Limited, Hampton, UK).
Data collection. Demographic data were collected from the hospitalrecords. The total number of breaths analyzed for each infant was noted,including their time after birth. Details of the waveforms of pressure, flow,and tidal volume were carefully analyzed to identify the breathing patterns.Breaths of each pattern were analyzed in detail and the following parametersnoted (see also Figures): respiratory rate, inspiration pattern and duration,expiratory hold (the time from zero flow at the end of inspiration to the startof the main expiratory flow), expiratory duration, postexpiratory pause (thetime from zero flow at the end of expiration to the start of positive flow at thebeginning of inspiration), duration of each breath, peak inspiratory flow, peakexpiratory flow; inspiratory and expiratory tidal volumes (15).
Data analysis. Data were analyzed with SPSS (SPSS for windows, version12.0, 2005, Chicago, IL) and presented as mean (SD) or number (%) whereappropriate.
RESULTS
Recordings from 103 infants were examined in detail, 91 ofthese were excluded from analysis for the following reasons:58 infants were ventilated with mask and bag at the start ofresuscitation, 27 recordings showed mask leak and 6 flowsignals were disturbed by secretions or movement of the mask.Twelve infants had recordings that were suitable for analysis.Their mean (SD) gestational age was 29 (1.9) weeks, birth weight1220 (412) g and median (range) of the 5 min Apgar score was8 (8,9). Eleven received prenatal steroids and seven were born by
Received March 3, 2008; accepted April 18, 2008.Correspondence: Arjan B. te Pas, M.D., Division of Newborn Services, Royal
Women’s Hospital Melbourne, 132 Grattan Street, Carlton, VIC 3053, Australia; e-mail:[email protected]
AtP is recipient of a Ter Meulen fund grant for working visits, Royal NetherlandsAcademy of Arts and Sciences, The Netherlands. Abbreviation: CPAP, continuous positive airway pressure
0031-3998/08/6403-0281PEDIATRIC RESEARCH
Vol. 64, No. 3, 2008
Copyright © 2008 International Pediatric Research Foundation, Inc.Printed in U.S.A.
281
A50
caesarean section. The mean time for the start of recording theanalyzed breaths was 159 s (range, 26–301) after birth. None ofthe infants were intubated in the delivery room.
A total of 792 breaths (range, 17–105 per infant) wereanalyzed. The pattern of inspiration was similar in all breathswith a duration of 0.36 (0.11) s. There were five differentpatterns of expiratory flow (Figs. 1–5). Three were character-ized by interruption or braking of expiration and were seen in79% (627/792) of the breaths. They prolonged the expiratorytime to 1.6 (1.1) s. The remaining two patterns were seen in21% (165/792) of the breaths. These showed no evidence ofinterruption or braking of expiration and had an expirationtime of 0.53 (0.13) s. The numbers of breaths of each patternfor each patient are shown in Table 1. In most infants, morethan one pattern was observed. The respiratory parameters ofeach pattern are shown in Table 2.
Expiratory hold. In 47% (372/792) of the breaths expirationwas braked to a complete hold, postponing the main expira-tory flow (Fig. 1). This pattern was characterized by a periodof no expiratory flow ending with a single expiratory flowpeak or multiple expiratory flow peaks. Expiration was im-mediately followed by an inspiration, i.e., there was no pos-texpiratory pause.
Slow expiration. In 10% (83/792) of breaths, the expiratorypattern was characterized by an initial low expiratory flow rateending with a single expiratory flow peak late in expirationand/or frequently interrupted expiratory flow waves (Fig. 2).These were immediately followed by an inspiration.
Crying/grunting. In 22% (172/792) of the breaths, expira-tion was slowed by either grunting or crying (Fig. 3A and B).This pattern was characterized by a large inspiration followed byhigh frequency interruptions to the expiratory flow wave, ob-
served as a noise signal in the wave. Expiration was then imme-diately followed by inspiration. We categorized crying and grunt-ing as one pattern as it was difficult in most cases to distinguishbetween them from the recording and we did not record sound.
Patterns where expiration was not braked. In 21% (165/792) of the breaths, an unbraked or normal expiratory patternwas seen (Figs. 4 and 5). This was characterized by uninter-rupted expiration with peak expiratory flow early in expira-tion. Expiration was not prolonged (I:E time approximately1:1.5) although sometimes the expiratory flow was followedby an expiratory pause before the next inspiration (Fig. 4). In91/165 of these breaths, the pattern was characterized by a
Figure 1. Two examples of the expiratory hold pattern, characterized by aperiod of no expiratory flow ending with a single expiratory flow peak ormultiple expiratory flow peaks. Expiration is immediately followed by aninspiration; there is no postexpiratory pause. *Note that the software settingis that it resets volume trace as soon as expiratory flow trace reaches baseline.
Figure 2. An example of a slow expiration pattern, expiratory flow is slowedor frequently interrupted. Expiration is immediately followed by an inspira-tion, there is no postexpiratory pause.
Figure 3. A, B. Two examples from patient 11 with the crying/gruntingpattern (scale of flow is doubled to fit peak inspiratory flow rate). In bothbreaths expiration is slowed down and a noise signal with high frequency isvisible. *Note that the software setting is that it resets volume trace as soonas expiratory flow trace reaches baseline.
282 TE PAS ET AL.
A51
respiratory rate greater than 60/min by shortening the expira-tory time (I:E time to approximately 1:1) and small tidalvolumes. In these breaths, there was no postexpiratory pausebefore inspiration. We named this the panting pattern (Fig. 5).
DISCUSSION
Most of the breaths of these very preterm infants, recordedin the minutes after birth, were characterized by interruption,or braking, of the expiratory flow. This slowed or even stoppedthe expiration and increased the expiratory time. Slow expiration,grunting/crying, and the expiratory hold pattern each representdifferent forms of expiratory braking (16–23). These patternswere shown to prevent a loss of lung volume during expiration inolder term (16–23) and preterm infants (19,20).
Airflow is controlled throughout the breathing cycle byreflex action of diaphragmatic and laryngeal muscle activity(24–30). These reflexes are present at birth in term andpreterm infants and have a crucial role in the control ofbreathing and lung volume (22,31–37). There are two mecha-nisms for stopping or slowing expiratory flow and maintaining anelevated lung volume during expiration. The first is diaphrag-matic postinspiratory activity, which slows the rate of lungdeflation by counteracting its passive recoil (16–20). The secondis by closure or narrowing of the larynx (19,21–23). The expira-tory noise associated with vocal cord adduction, i.e., grunting hasbeen recognized as a dynamic braking of expiration by the larynx(22,38). In absence of an audible grunt, the larynx still maintainsa dynamic role in slowing expiration (27).
The pattern that we most frequently observed, the expira-tory hold pattern, has similar characteristics to the firstbreaths observed in term infants at birth (13,39,40). Afterinspiration, gas is held in the lung under pressure by laryngealadduction braking the expiratory flow. This has also beenobserved in preterm and term infants during spontaneousbreathing later in life, but did not occur as frequently as in ourstudy (16–19,22,23,38).
During braked expiration, the closed or narrowed glottis,with increased intra-pulmonary pressure from abdominal mus-cle contraction, causes the airway pressure to be maintainedabove atmospheric. This could represent an important forcefor clearing the fluid from the lung, facilitating distribution ofgas within the lung, and splinting the alveoli and airways open(9,10,13,14,22,39).
Another strategy infants can use to increase end-expiratoryvolume is a high respiratory rate. This was seen in the pantingpattern. The shortened expiratory time reduces the time for gasto leave the lung (35).
The spontaneous gas flow patterns we observed are differentfrom those seen when manual inflations are given duringneonatal resuscitation (Fig. 6). The neonatal resuscitationguidelines do not mandate any particular ventilation pattern,although they do suggest a rate of 40–60/min (41). Prolongedinflations are mentioned but not mandated for the initialbreaths (41). This study demonstrates that very preterm infantshave specific mechanisms to inflate their lungs and try to keepthem inflated immediately after birth. It is possible that asimilar strategy used during the initial resuscitation of apreterm infant who is apnoeic or breathes insufficiently, i.e.,inflation followed by a hold and a short expiration, may bemore effective than traditional techniques which use a shortinspiratory time and no hold.
Making recordings of how very preterm infants breatheimmediately after birth is very difficult. Because of the con-straints of not interfering with the infant’s care or resuscitationand yet applying a facemask with a pneumotachograph andCPAP device attached, as soon as possible after birth, limitedthe number of infants that could be recorded in the timeavailable and the number of recordings that could be made. Thisis an area of research where it is not possible to study largenumbers in detail and obtaining a recording of sufficient qualityfor analysis limited the number of inflations that could be studied.In particular, it caused a large variation in the number of breaths
Figure 4. An example of a normal expiratory pattern without braking,characterized by a peak expiratory flow† at the beginning of expiration andshowing a postexpiratory pause*.
Figure 5. An example of panting pattern with respiratory rate of 96/min,tidal volume 2 mL/kg and expiratory time is approximately equal toinspiratory time.
283BREATHING OF PRETERM INFANTS AT BIRTH
A52
analyzed per infant. Analyzing a longer breath sequence over alonger period would have been ideal but was not practical. Futurestudies should endeavor to achieve this. Similar studies immedi-ately after birth done by Karlberg et al. (39), Milner and Sauders(9), and Mortola et al. (13) also recorded a limited number ofbreaths in small groups of patients.
Our observation could be biased by the practical difficultiesof making respiratory recordings immediately after birth. Our
neonatal resuscitation research team has extensive experiencein making physiologic recordings and recognized it was verydifficult to get long recordings without ventilation, mask leak,movement of the mask or patient, secretions, etc. All record-ings with artifacts have been excluded from analysis and asexpected, only a minority of recordings was suitable forinclusion in this study. Both small number of recordings andthe wide range of analyzed breaths per infant could have ledto a bias. Therefore, we cannot assume that the patternsobserved in this study represent the patterns of all spontane-ously breathing preterm infants. Nevertheless, they are similarto those described in newly born term infants (13,39,40) andin older preterm infants (19,20).
Although a mask on the face is a standard procedure fordelivering positive airway pressure to infants after birth it mightinfluence the breathing pattern by stimulating respiratory reflexesand thereby influence the tidal volume and respiratory rate(42,43). However, it is impossible to measure inspiratory andexpiratory flows and tidal volumes in this situation without use offace mask (44,45). Previous studies reported no adverse effects ofa face mask during the first breaths (7–14). It is possible that theadded dead space of the face mask caused carbon dioxide accu-mulation and influenced the breathing pattern (46,47). However,the wide diameter of the mask reduced rebreathing and firmapplication of the mask reduced the dead space (15).
In these very preterm infants, it was not appropriate to use afacemask without also providing CPAP. The use of CPAP mighthave influenced the breathing pattern. However, the patterns wehave recorded are similar to some breathing patterns described interm infants at birth without CPAP (13,39,40).
Figure 6. Part of a recording of resuscitation of a neonate born at 26 wkgestation showing two inflations with Laerdal mask and bag. Inspiration isimmediately followed by expiration, with no expiratory hold or pause.
Table 2. Respiratory parameters for each breathing patterns for all infants
Parameter mean (SD)Expiratory hold
n " 372
Slowexpiration
n " 83Crying/grunting
n " 172
Unbraked expiration, normalrespiratory rate
n " 84Pantingn " 81
Respiratory rate (min#1) 32 (11) 48 (16) 42 (18) 54 (4) 88 (18)Inspiratory time (sec) 0.36 (0.10) 0.34 (0.15) 0.38 (0.14) 0.40 (0.08) 0.34 (0.07)Expiratory hold (sec) 1.53 (1.10) 0 0 0 0Expiratory time (sec)* 1.85 (1.14) 1.10 (0.90) 1.30 (0.75) 0.65 (0.11) 0.41 (0.14)Postexpiratory pause (sec) 0 0 0 0.07 (0.03) 0Peak inspiratory flow (ml/s) 32 (16) 29 (130) 63 (37) 22 (9) 20 (10)Peak expiratory flow (ml/s) #42 (30) #24 (14) #36 (24) #19 (15) #18 (12)Tidal volume (ml/kg) 5.8 (4.1) 3.5 (2.3) 7.5 (4.2) 4.2 (1.5) 3.1 (1.7)
* Expiratory time of expiratory hold pattern is time of expiration plus duration of the hold.
Table 1. Number of each type of breathing pattern for each infant
Infant Gestation (wk) Expiratory hold Crying/grunting Slow expiration Panting Unbraked expiration, normal respiratory rate Total
1 26 49 10 5 2 2 682 28 22 0 0 0 0 223 28 32 5 1 0 0 384 30 8 0 1 31 2 425 29 37 5 1 12 0 556 29 58 28 10 5 3 1047 27 46 0 10 18 15 898 27 5 11 0 11 51 789 32 42 16 37 2 1 9810 29 56 22 17 10 0 10511 32 10 66 0 0 0 7612 28 7 9 1 0 0 17Total 372 172 83 91 74 792
284 TE PAS ET AL.
A53
A nCPAP of 8 cm H2O was used for several reasons. Studiesshow a distending pressure is important for maintaining FRC(48), increasing lung compliance and improving oxygenation,and that 8 cm H2O is more effective than a lower pressure (49).Studies have used pressures up to 10 cm H2O (50) and Gregoryet al.(51) used a nCPAP pressure up to 12 mm Hg. Immediatelyafter birth is the time when most assistance is needed todevelop and maintain an FRC because the lungs are fluidfilled and the lungs are stiff. It is likely that a relatively highpressure may be necessary in the first few minutes. Animalstudies show that a pressure of 8 cm H2O improves oxy-genation and reduces lung injury better than lower pres-sures (52,53). The optimum nCPAP pressure for treatingindividual very preterm infants from birth is unknown.
In conclusion, this is the first study reporting the spontane-ous breathing patterns in very preterm infants in the minutesafter birth. We frequently observed prolongation of expirationin very preterm infants treated with facemask CPAP immediatelyafter birth, predominantly characterized by a breath hold. Thisbraking of expiration is recognized as an attempt to defend lungvolume. Our speculation is that techniques of positive pressureventilation that mimic expiratory braking might improve theeffectiveness of respiratory support in the delivery room.
REFERENCES
1. Gerhardt T, Bancalari E 1980 Chestwall compliance in full-term and prematureinfants. Acta Paediatr Scand 69:359–364
2. Heldt GP, McIlroy MB 1987 Dynamics of chest wall in preterm infants. J ApplPhysiol 62:170–174
3. Barker PM, Gowen CW, Lawson EE, Knowles MR 1997 Decreased sodium ionabsorption across nasal epithelium of very premature infants with respiratory distresssyndrome. J Pediatr 130:373–377
4. Zelenina M, Zelenin S, Aperia A 2005 Water channels (aquaporins) and their role forpostnatal adaptation. Pediatr Res 57:47R–53R
5. Aly H, Massaro AN, Patel K, El Mohandes AA 2005 Is it safer to intubate prematureinfants in the delivery room? Pediatrics 115:1660–1665
6. Thomson MA 2002 Continuous positive airway pressure and surfactant; combineddata from animal experiments and clinical trials. Biol Neonate 81:16–19
7. Karlberg P, Koch G 1962 Respiratory studies in newborn infants. III. Developmentof mechanics of breathing during the first week of life. A longitudinal study. ActaPaediatr Suppl 135:121–129
8. Karlberg P 1960 The adaptive changes in the immediate postnatal period, withparticular reference to respiration. J Pediatr 56:585–604
9. Milner AD, Sauders RA 1977 Pressure and volume changes during the first breathof human neonates. Arch Dis Child 52:918–924
10. Vyas H, Milner AD, Hopkins IE 1981 Intrathoracic pressure and volume changesduring the spontaneous onset of respiration in babies born by cesarean section andby vaginal delivery. J Pediatr 99:787–791
11. Vyas H, Milner AD, Hopkin IE, Falconer AD 1983 Role of labour in the establish-ment of functional residual capacity at birth. Arch Dis Child 58:512–517
12. Vyas H, Field D, Milner AD, Hopkin IE 1986 Determinants of the first inspiratoryvolume and functional residual capacity at birth. Pediatr Pulmonol 2:189–193
13. Mortola JP, Fisher JT, Smith JB, Fox GS, Weeks S, Willis D 1982 Onset ofrespiration in infants delivered by cesarean section. J Appl Physiol 52:716–724
14. Saunders RA, Milner AD 1978 Pulmonary pressure/volume relationships during thelast phase of delivery and the first postnatal breaths in human subjects. J Pediatr93:667–673
15. Stick S 1996 Measurements during tidal breathing. In: Stocks J, Sly PD, Tepper RS,Morgan WJ (eds) Infant Respiratory Function Testing. New York: John Wiley &Sons Inc, pp 118–119
16. Kosch PC, Stark AR 1984 Dynamic maintenance of end-expiratory lung volume infull-term infants. J Appl Physiol 57:1126–1133
17. Lopes J, Muller NL, Bryan MH, Bryan AC 1981 Importance of inspiratory muscletone in maintenance of FRC in the newborn. J Appl Physiol 51:830–834
18. Mortola JP, Milic-Emili J, Noworaj A, Smith B, Fox G, Weeks S 1984 Musclepressure and flow during expiration in infants. Am Rev Respir Dis 129:49–53
19. Radvanyi-Bouvet MF, Monset-Couchard M, Morel-Kahn F, Vicente G, Dreyfus-Brisac C 1982 Expiratory patterns during sleep in normal full-term and prematureneonates. Biol Neonate 41:74–84
20. Stark AR, Cohlan BA, Waggener TB, Frantz ID III, Kosch PC 1987 Regulation ofend-expiratory lung volume during sleep in premature infants. J Appl Physiol62:1117–1123
21. Milner AD, Saunders RA, Hopkin IE 1978 Is air trapping important in the mainte-nance of the functional residual capacity in the hours after birth? Early Hum Dev2:97–105
22. Fisher JT, Mortola JP, Smith JB, Fox GS, Weeks S 1982 Respiration in newborns:development of the control of breathing. Am Rev Respir Dis 125:650–657
23. Mortola JP, Magnante D, Saetta M 1985 Expiratory pattern of newborn mammals.J Appl Physiol 58:528–533
24. Kosch PC, Hutchinson AA, Wozniak JA, Carlo WA, Stark AR 1988 Posteriorcricoarytenoid and diaphragm activities during tidal breathing in neonates. J ApplPhysiol 64:1968–1978
25. Kosch PC, Davenport PW, Wozniak JA, Stark AR 1985 Reflex control of expiratoryduration in newborn infants. J Appl Physiol 58:575–581
26. England SJ, Kent G, Stogryn HA 1985 Laryngeal muscle and diaphragmaticactivities in conscious dog pups. Respir Physiol 60:95–108
27. Harding R, Johnson P, McClelland ME 1980 Respiratory function of the larynx indeveloping sheep and the influence of sleep state. Respir Physiol 40:165–179
28. Farber JP 1978 Laryngeal effects and respiration in the suckling opossum. RespirPhysiol 35:189–201
29. Henderson-Smart DJ, Johnson P, McClelland ME 1982 Asynchronous respiratoryactivity of the diaphragm during spontaneous breathing in the lamb. J Physiol327:377–391
30. MacFarlane PM, Frappell PB, Mortola JP 2002 Mechanics of the respiratory systemin the newborn tammar wallaby. J Exp Biol 205:533–538
31. Paintal AS 1969 Mechanism of stimulation of type J pulmonary receptors. J Physiol203:511–532
32. Coleridge HM, Coleridge JC 1994 Pulmonary reflexes: neural mechanisms ofpulmonary defense. Annu Rev Physiol 56:69–91
33. Widdicombe JG 1964 Respiratory reflexes. In: Fenn W, Rahn H (eds) Handbook ofPhysiology, Section 3, Respiration. Vol. 2. Washington, DC: American Physiolog-ical Society, pp 585–630
34. Cross KW, Klaus M, Tooley WH, Weisser K 1960 The response of the new-bornbaby to inflation of the lungs. J Physiol 151:551–565
35. Olinsky A, Bryan MH, Bryan AC 1974 Influence of lung inflation on respiratorycontrol in neonates. J Appl Physiol 36:426–429
36. Hannam S, Ingram DM, Rabe-Hesketh S, Milner AD 2001 Characterisation of theHering-Breuer deflation reflex in the human neonate. Respir Physiol 124:51–64
37. Hannam S, Ingram DM, Milner AD 1998 A possible role for the Hering-Breuerdeflation reflex in apnea of prematurity. J Pediatr 132:35–39
38. Lindroth M, Johnson B, Ahlstrom H, Svenningsen NW 1981 Pulmonary mechanicsin early infancy. Subclinical grunting in low-birth-weight infants. Pediatr Res15:979–984
39. Karlberg P, Cherry RB, Escardo FE, Koch G 1962 Pulmonary ventilation andmechanics of breathing in the first minutes of life, including the onset of respiration.Acta Paediatr Scand 51:121–136
40. Engstrom L, Karlberg P, Rooth G, Tunell R 1966 The Onset of Respiration, a Studyof Respiration and Changes in Blood Gases and Acid-Base Balance. New York:Association for the aid of crippled children, pp 17–23
41. The International Liaison Committee on Resuscitation 2006 The InternationalLiaison Committee on Resuscitation (ILCOR) consensus on science with treatmentrecommendations for pediatric and neonatal patients: neonatal resuscitation. Pedi-atrics 117:e978–e988
42. Fleming PJ, Levine MR, Goncalves A 1982 Changes in respiratory pattern resultingfrom the use of a facemask to record respiration in newborn infants. Pediatr Res16:1031–1034
43. Dolfin T, Duffty P, Wilkes D, England S, Bryan H 1983 Effects of a face maskand pneumotachograph on breathing in sleeping infants. Am Rev Respir Dis128:977–979
44. Stocks J 1999 Lung function testing in infants. Pediatr Pulmonol Suppl 18:14–2045. Schmalisch G, Foitzik B, Wauer RR, Stocks J 2001 Effect of apparatus dead space
on breathing parameters in newborns: “flow-through” versus conventional tech-niques. Eur Respir J 17:108–114
46. Marsh MJ, Ingram D, Milner AD 1993 The effect of instrumental dead space onmeasurement of breathing pattern and pulmonary mechanics in the newborn. PediatrPulmonol 16:316–322
47. Morris MG 1999 A simple new technique to measure the effective dead space of theface mask with a water volumeter in infants. Eur Respir J 14:1163–1166
48. Thome U, Topfer A, Schaller P, Pohlandt F 1998 The effect of positive endexpira-tory pressure, peak inspiratory pressure, and inspiratory time on functional residualcapacity in mechanically ventilated preterm infants. Eur J Pediatr 157:831–837
49. Elgellab A, Riou Y, Abbazine A, Truffert P, Matran R, Lequien P, Storme L 2001Effects of nasal continuous positive airway pressure (NCPAP) on breathing patternin spontaneously breathing premature newborn infants. Intensive Care Med27:1782–1787
50. Kamper J, Ringsted C 1990 Early treatment of idiopathic respiratory distresssyndrome using binasal continuous positive airway pressure. Acta Paediatr Scand79:581–586
51. Gregory GA, Kitterman JA, Phibbs RH, Tooley WH, Hamilton WK 1971 Treatmentof the idiopathic respiratory-distress syndrome with continuous positive airwaypressure. N Engl J Med 284:1333–1340
52. Probyn ME, Hooper SB, Dargaville PA, McCallion N, Harding R, Morley CJ 2005Effects of tidal volume and positive end-expiratory pressure during resuscitation ofvery premature lambs. Acta Paediatr 94:1764–1770
53. Mulrooney N, Champion Z, Moss TJ, Nitsos I, Ikegami M, Jobe AH 2005 Surfactantand physiological responses of preterm lambs to continuous positive airway pres-sure. Am J Respir Crit Care Med 171:488–493
285BREATHING OF PRETERM INFANTS AT BIRTH
A54
Breathing Patterns in Preterm and Term Infants ImmediatelyAfter Birth
ARJAN B. TE PAS, CONNIE WONG, C. OMAR F. KAMLIN, JENNIFER A. DAWSON, COLIN J. MORLEY, AND PETER G. DAVIS
Division of Newborn Services, Royal Women’s Hospital, Carlton, Victoria 3053, Australia
ABSTRACT: There is limited data describing how preterm and terminfants breathe spontaneously immediately after birth. We studiedspontaneously breathing infants !29 wk immediately after birth.Airway flow and tidal volume were measured for 90 s using a hotwire anemometer attached to a facemask. Twelve preterm and 13term infants had recordings suitable for analysis. The median (inter-quartile range) proportion of expiratory braking was very high inboth groups (preterm 90 !74–99" vs. term 87 !74–94"%; NS). Cryingpattern was the predominant breathing pattern for both groups (62!36–77"% vs. 64 !46–79"%; NS). Preterm infants showed a higherincidence of expiratory hold pattern (9 !4–17"% vs. 2 !0–6"%; p #0.02). Both groups had large tidal volumes (6.7 !3.9" vs. 6.5 !4.1"mL/kg), high peak inspiratory flows (5.7 !3.8" vs. 8.0 !5" L/min),lower peak expiratory flow (3.6 !2.4" vs. 4.8 !3.2" L/min), shortinspiration time (0.31 !0.13" vs. 0.32 !0.16" s) and long expirationtime (0.93 !0.64" vs. 1.14 !0.86" s). Directly after birth, both pretermand term infants frequently brake their expiration, mostly by crying.Preterm infants use significantly more expiratory breath holds todefend their lung volume. (Pediatr Res 65: 352–356, 2009)
Between 1960 and 1986, observational data were gatheredon breathing patterns immediately after birth from small
numbers of term infants and were used to inform internationalguidelines for neonatal resuscitation (1–8). These studiesdemonstrated that the first breaths tend to be deeper and longerthan subsequent breaths and are characterized by a short deepinspiration followed by a prolonged expiratory phase. This isknown as expiratory braking and helps to develop and main-tain functional residual capacity (FRC) during the immediatenewborn period when the lung is partially liquid filled and thechest wall is very compliant (9,10). Although this respiratorypattern has also been observed in preterm and term infantslater in life (11–15), there are no data describing the breathingpattern of very preterm infants immediately after birth.
Antenatal glucocorticoid treatment has greatly improvedpostnatal lung function and many preterm infants breathe welland establish an FRC at birth by themselves or with only thesupport of continuous positive airway pressure (16–20). How-ever, preterm infants have a poor respiratory drive, weakmuscles, flexible ribs, surfactant deficiency, and impaired lungliquid clearance, which make it difficult for them to breatheeasily at birth (21–24). With these inherent problems, we
hypothesized that preterm infants in the minutes after birthwill show more expiratory braking than term infants.
The aim of this study was to compare the breathing patternsof preterm and term infants immediately after birth.
METHODS
All inborn infants, term and preterm !29 wk gestation, who were expectedto require no respiratory support at birth, were eligible for this study. Thestudy was approved by the Royal Women’s Hospital Research and Ethicscommittees. Written consent was obtained before birth.
Immediately after birth, as soon as the infant was placed on the resusci-tation trolley, a facemask (Laerdal round mask, Laerdal Stavanger, Norway)was applied to the face, enclosing the mouth and nose. To ensure that therewas no mask leak, a finger was applied around the infant’s chin and heldfirmly during the recording. A hot wire anemometer (Florian: AcutronicMedical Systems AG, Zug, Switzerland) was attached proximally to theLaerdal mask measuring inspiratory and expiratory gas flow (see also Fig. 1)(25). The flow signal was integrated to provide inspired and expired tidalvolume. The signals of airway flow and tidal volumes were digitized andrecorded at 200 Hz using a neonatal respiratory physiologic recording pro-gram (Spectra, Grove Medical Limited, Hampton, UK).
To eliminate the risk of carbon dioxide retention caused by the mask (26),a bias flow of 2 L/min of air was fed in through the face mask and theanemometer rezeroed (Fig. 1) (26). The dead space of the hot wire anemom-eters is 1 mL and clinically negligible (27).
To minimize interference with the normal monitoring and stabilizationof preterm and term infants at birth, we recorded for only 90 s. If therewere signs of respiratory compromise, the study was abandoned andventilatory support was given according to the Australian Neonatal Re-suscitation guidelines (28).
Data collection. The following clinical data were collected: gestationalage, birth weight, sex, mode of delivery, and Apgar score. The total numberof breaths analyzed for each infant was noted, including their time after birth.Details of the waveforms of pressure, flow, and tidal volume were carefullyanalyzed to identify breathing patterns. Breaths of each pattern were analyzedin detail by a breath-to-breath analysis and the following parameters werecalculated: respiratory rate, inspiration pattern and duration, expiratory hold(the time from zero flow at the end of inspiration to the start of the mainexpiratory flow), expiratory duration, postexpiratory pause (the time fromzero flow at the end of expiration to the start of positive flow at the beginningof inspiration), duration of each breath, peak inspiratory flow, peak expiratoryflow; inspiratory and expiratory tidal volumes (29).
Recordings were excluded if: a) there were signs of mask leak, b) the flowsignal was disturbed by secretions or infant movements, c) there were signs ofmovement of the mask, or if ventilatory support was given.
Based on our earlier observations of spontaneous breathing patterns ininfants at birth, we divided the breathing patterns by the type of expiration:braked or unbraked. These patterns were defined as follows.
Expiratory hold. Expiration is braked to a complete hold, postponing themain expiratory flow (Fig. 2) . This pattern is characterized by a period of noexpiratory flow ending with a single expiratory flow peak or multiple expi-ratory flow peaks. Expiration is immediately followed by an inspiration, i.e.,there is no postexpiratory pause.
Slow expiration. Expiration is characterized by an initial low expiratoryflow rate ending with a single expiratory flow peak late in expiration and/or
Received May 6, 2008; accepted September 24, 2008.Correspondence: Arjan B. te Pas, M.D., Division of Newborn Services, Royal
Women’s Hospital Melbourne, 132 Grattan Street, Carlton, Victoria 3053, Australia;e-mail: [email protected]
A.B.P. and J.A.D. are recipients of a Royal Women’s Hospital postgraduate scholarship. Abbreviations: FRC, functional residual capacity
0031-3998/09/6503-0352PEDIATRIC RESEARCH
Vol. 65, No. 3, 2009
Copyright © 2009 International Pediatric Research Foundation, Inc.Printed in U.S.A.
352
A55
frequently interrupted expiratory flow waves (Fig. 3). These are immediatelyfollowed by an inspiration.
Crying. Expiration is slowed by crying (Fig. 4). This pattern is character-ized by a large inspiration followed by high frequency interruptions to theexpiratory flow wave, which can be seen in the wave as a noise signal.Expiration is then immediately followed by inspiration. A cry was differen-tiated from a grunt in the recording by the following: 1) during the recordingsession, we labeled the breath in the recording the moment a cry or a gruntwas heard, 2) in general, the noise signal in the expiratory pattern of a cry hasa larger amplitude and higher frequency than a grunt.
Grunting. Expiration is slowed by grunting (Fig. 5). This pattern ischaracterized by a large inspiration followed by high frequency interruptionsto the expiratory flow wave, which can be seen in the wave as a noise signal.Expiration is then immediately followed by inspiration.
Unbraked expiration patterns. These are characterized by uninterruptedexpiration with peak expiratory flow early in expiration. Expiration is not
prolonged (I:E time approximately 1:1.5) although sometimes the expiratoryflow is followed by an expiratory pause before the next inspiration (Fig. 6).
The unbraked breathing pattern was called panting when the respiratoryrate was greater than 60/min. This is achieved by shortening the expiratorytime (I:E time to approximately 1:1) and often small tidal volumes werenoted. In these breaths, there was no postexpiratory pause before inspiration(Fig. 7).
Figure 1. Placement of the open facemask demonstrated on a mannequin.The facemask was applied to the face, enclosing the mouth and nose. Toensure that there was no mask leak a finger was applied around the infant’schin was held firmly during the recording. A hot wire anemometer wasattached proximally to the Laerdal mask. A bias flow of 2 L/min of air wasgiven through the facemask.
Figure 2. An example of three breaths showing an expiratory hold pattern ina spontaneously breathing infant of 31 wk. The pattern is characterized by aperiod of no expiratory flow ending with a single expiratory flow peak ormultiple expiratory flow peaks. Expiration is immediately followed by aninspiration; there is no postexpiratory pause. Note in all figures that thesoftware resets the volume trace at the end of inspiration and end of expirationas soon as the flow reaches baseline. In this example, more volume enters thanleaves the lung, which implies an increase in functional residual capacity.
Figure 3. An example of a breath with a slow expiration pattern in an infantof 30 wk gestation. In this pattern, expiratory flow is slowed or frequentlyinterrupted. Expiration is immediately followed by an inspiration, i.e., there isno postexpiratory pause.
Figure 4. Examples of crying pattern in a term infant of 40 wk gestation anda preterm infant of 29 wk gestation. This pattern is characterized by a largeinspiration followed by high frequency interruptions to the expiratory flowwave, which can be seen in the wave as a noise signal. Expiration is thenimmediately followed by inspiration. A cry has a higher amplitude andfrequency than a grunt.
Figure 5. An example of expiration slowed by grunting in an infant of 33 wkof gestation. This pattern is characterized by a large inspiration followed byhigh frequency interruptions to the expiratory flow wave, which can be seenin the wave as a noise signal. Expiration is then immediately followed byinspiration. In general, a grunt has lower amplitude and frequency than a cry.
353TE PAS ET AL.
A56
Data analysis. Data are presented as numbers and proportions (%) forcategorical variables, and means with SD for normally distributed continuousvariables and median (interquartile range) when the distribution was skewed. The"2 test was used to detect differences between the two groups for categoricalvariables and student’s t test for continuous variables. Data were analyzed usingSPSS software (SPSS for windows, version 15.0, 2006, Chicago, IL).
RESULTS
During the period August 2007–February 2008 there were62 eligible infants. In seven infants, the mothers were notapproached because the midwife/doctor considered it wouldbe too stressful for the parents. In ten infants, the parents didnot consent. In four infants, no recording was obtained due totechnical problems. Forty-one recordings were made in 27preterm infants and 14 term infants. Of these 16 recordingswere excluded from analysis because secretions causing dirtyflow signal in one term and one preterm infant and a further 14preterm infants needed ventilatory support. Thus, there were25 recordings available for analysis, 12 preterm and 13 terminfants. We analyzed 769 breaths in the preterm infants and749 breaths in the term infants. The characteristics of theinfants studied are shown in Table 1.
Patterns of Breathing. The incidence of breaths of eachpattern for both groups are shown in Table 2, and examples ofeach pattern are shown in Figures 2–7.
The median (interquartile range) proportion of expiratorybraking patterns was very high in both groups, in preterminfants 90 (74–99)% and in term infants 87 (74–94)% (p #0.6). Of the braked patterns, expiratory hold pattern occurredmore often in preterm infants (9 !4–17"% vs. 2 !0–6"%; p #0.02). The most common breathing pattern for both groupswas the crying pattern (62 !36–77"% vs. 64 !46–79"%; p #0.4). The incidence of unbraked pattern was low in bothgroups (10 !3–27"% vs. 13 !6–25"%; p # 0.6).
Respiratory Parameters. There was no difference in theaverage tidal volume measured between preterm and terminfants (6.7 !3.9" vs. 6.5 !4.1" mL/kg; p # 0.5). The breaths inboth groups were characterized by a high peak flow duringinspiration (5.7 !3.8" vs. 8.0 !5" mL/kg; p $ 0.001) and alower peak flow during expiration (3.6 !2.4" vs. 4.8 !3.2"mL/kg; p $ 0.001). The average inspiration time was shortand not different between the groups (0.31 !0.13" vs. 0.32!0.16" s; p # 0.5). The average expiration time in both groupswas long, but in preterm infants, it was shorter than in terminfants (0.93 !0.64" vs. 1.14 !0.86" s; p $ 0.001).
The values of the respiratory parameters of both groupswhen divided into breaths with braked and unbraked expira-tion are shown in Table 3. In both groups, breaths with brakedexpiration were characterized by larger tidal volume, largerpeak inspiratory flow, longer expiration time, and smallerrespiratory rate (see Table 3).
DISCUSSION
We have shown in this study that both term and preterminfants frequently brake their expiration in the first minutes oflife. This is achieved most commonly by crying. The cryingpattern has not been described in earlier reports (7,30,31). Allother interrupted expiratory patterns shown in this study weredescribed previously and were considered strategies to defend
Figure 6. An example of a breath with a “normal” expiratory pattern withoutbraking in an infant of 32 wk gestation. This pattern is characterized by a peakexpiratory flow at the beginning of expiration and sometimes showing apostexpiratory pause.
Figure 7. Example of panting pattern in a term infant. This pattern ischaracterized by a respiratory rate greater than 60/min (in this example130/min) by shortening the expiratory time (I:E time to approximately 1:1)and small tidal volumes are noted (in this example 2.5–3.0 mL/kg). In thesebreaths, there is no postexpiratory pause before inspiration.
Table 1. Characteristics of the term and preterm infants
CharacteristicsPreterm
(N # 12)Term
(N # 13) p
Gestational age, wk, mean(SD)
32 (2.2) 38.9 (0.9) $0.001
Birth weight, g, mean (SD) 2000 (560) 3340 (530) $0.001Apgar at 5 min, median (IQR) 9 (9–9) 9 (8–9) NSTime start recording, s, mean
(SD)30 (15) 29 (14) NS
Caesarean delivery (%) 33 54 NS
Table 2. The proportion of breaths with different expiratorypatterns for preterm and term infants
12 Preterm 13 Term p
No. breaths 58 (45–82) 62 (34–73) 0.5Braked expiration
Prolonged 12 (3–17)% 11 (3–18)% 0.9Hold 9 (4–17)% 2 (0–6)% 0.02Grunt 2 (0–10)% 0 (0–6)% 0.5Cry 62 (36–77)% 64 (46–79)% 0.4
Unbraked expirationNormal 4 (0–16)% 11 (3–18)% 0.9Panting 0 (0–1)% 0 (0–12)% 0.5
354 BREATHING PATTERNS AT BIRTH
A57
lung volume (11,12,14,15,32–35). We were not able to showa difference in breathing between the groups, except thatpreterm infants use the expiratory breath hold more often thanterm infants.
This is the first study comparing the breathing patterns ofterm and preterm infants immediately after birth. Karlberg etal. (30) and Mortola et al. (7) reported the first few breaths interm infants and described similar interruptions in expirationwith small or zero flow. These expiratory interruptions orbraking were also observed in term infants at 10 min of life(11). They also occur in preterm and term infants later in lifeduring active sleep (15,33).
In newborns, with a very compliant chest wall, it is likelythat expiratory braking mechanisms help maintain FRC. Thereare two mechanisms for stopping or slowing expiratory flowand maintaining an elevated lung volume during expiration.Diaphragmatic postinspiratory activity slows the rate of lungdeflation by counteracting its passive recoil (12,14,15,32,33).Closure or narrowing of the larynx increases the resistance toexpiration (11,15,34,35). During braked expiration, the closedor narrowed glottis, with increased intrapulmonary pressurefrom abdominal muscle contraction, causes the airway pres-sure to be maintained above atmospheric. This helps clearfluid from the lung, facilitate distribution of gas within thelung, and splint the alveoli and airways open (3,4,7,8,11,30).
A cry at birth reassures clinicians that the infant is capableof taking a deep inspiration (36). The sound is producedduring forced expiration by vibration of the vocal cords (37).In this study, we have shown that crying is a similar breathingpattern to grunting. It is an interruption of the expiratory flowwith a higher frequency (Fig. 3) and amplitude than grunting.Grunting has been recognized as a dynamic braking of expi-ration by the larynx (11,13), but crying has not been describedearlier as a braking mechanism. The most likely reason for thisis that in earlier studies breaths were excluded from analysis ifthe infant was crying (7,11). Interestingly, Engstrom et al.recorded sound and intraesophageal pressure in the first min-utes of life. Although they did not record flow and volume,they described breathing in the first minutes of life as a“crying period” characterized by deep breaths (31). It is likelythat the prolonged high intrathoracic pressure on expirationduring crying represents an important force for clearing fluidfrom the lung and facilitating lung aeration.
The tidal volumes in the nonbraked patterns are signifi-cantly lower in preterm infants compared with term infants.Preterm infants have difficulties aerating their lungs because
of poor respiratory drive, weak muscles, flexible ribs, surfac-tant deficiency, and impaired lung liquid clearance (21,22,24).We speculate that these factors become more obvious whenbreaths are not braked.
We observed an expiratory hold pattern more often inpreterm infants. Preterm infants may have difficulty aeratingtheir lungs and keeping them open and may need theseexpiratory hold maneuvers more often. It is possible that whenthese patterns are observed very frequently, it could indicatethe infant may need respiratory support or, when alreadystarted, need to be optimized.
The braked pattern was characterized in both groups by alarger tidal volume. This probably reflects the large proportionof crying pattern that is accompanied by a large inspiration.The peak inspiratory flow we measured during the cryingpattern in term infants is comparable to what has been mea-sured in crying term infants later in life (36). However, wemeasured a much lower peak expiratory flow, which couldindicate more braking occurred at birth than during cryinglater in life (36).
Making respiratory measurements immediately after birth isvery difficult. This is an area of research where it is notpossible to study large numbers in detail. We were only ableto record more mature preterm infants because we aimed torecord only those who were breathing without assistance. Tohave minimal interference with the monitoring and stabiliza-tion of preterm and term infants at delivery, we recorded foronly 90 s. It is possible that more differences between thegroups would have appeared if more immature preterm infantswere studied or if recordings were made for longer periods.
Although using a face mask with a pneumotachometerattached is an accepted and only method for measuring respi-ratory parameters after birth (37,38), it may have influencedthe breathing pattern by stimulating respiratory reflexes andthereby influencing the tidal volume and respiratory rate(39,40). In previous studies, no adverse effects of using a facemask during the first breaths were reported (1,2,7,11). Inaddition, an open face mask with a small pneumotachographattached causes less interference to the infants’ breathing thanmethods used in previous studies (1–6,8).
In conclusion, this is the first report that describes in detailthe breathing patterns of preterm and term infants immediatelyafter birth. Both preterm and term infants frequent brakeexpiration, most often represented by a crying pattern. Thecrying pattern has not been described before but seems to bea method of breathing that uses expiratory braking and facil-
Table 3. The respiratory parameters of the braked and unbraked breaths for both groups
Respiratory parameters, mean(SD)
Braked expiration Unbraked expiration
Preterm(N # 652 breaths)
Term(N # 598 breaths) p
Preterm(N # 117 breaths)
Term(N # 151 breaths) p
Tidal volume (mL/kg) 7.2 (3.8)* 6.8 (4.2)* NS 3.7 (2.2)* 5.5 (3.4)* $0.001Inspiration time (s) 0.32 (0.14) 0.33 (0.16) NS 0.30 (0.09) 0.30 (0.13) NSExpiration time (s) 1.03 (0.84)* 1.33 (1.02)* $0.001 0.41 (0.16)* 0.43 (0.26)* NSPeak flow inspiration (L/min) 6.2 (3.9)* 8.4 (5.2)* $0.001 2.9 (1.8)* 6.5 (4.2)* $0.001Peak flow expiration (L/min) %3.8 (2.4)* %4.3 (2.8)* $0.001 %3.0 (2.4)* %6.3 (4.0)* $0.001Respiration rate (min%1) 60 (30)* 50 (23)* $0.001 90 (26)* 91 (31)* NS
* Values of braked vs. unbraked pattern were significantly different in both groups (p $ 0.001).
355TE PAS ET AL.
A58
itates lung volume recruitment. We have shown that preterminfants use significantly more expiratory breath holds to de-fend their lung volume.
REFERENCES
1. Karlberg P, Koch G 1962 Respiratory studies in newborn infants. III. Developmentof mechanics of breathing during the first week of life. A longitudinal study. ActaPaediatr Suppl 135:121–129
2. Karlberg P 1960 The adaptive changes in the immediate postnatal period, withparticular reference to respiration. J Pediatr 56:585–604
3. Milner AD, Sauders RA 1977 Pressure and volume changes during the first breathof human neonates. Arch Dis Child 52:918–924
4. Vyas H, Milner AD, Hopkins IE 1981 Intrathoracic pressure and volume changesduring the spontaneous onset of respiration in babies born by cesarean section andby vaginal delivery. J Pediatr 99:787–791
5. Vyas H, Milner AD, Hopkin IE, Falconer AD 1983 Role of labour in the establish-ment of functional residual capacity at birth. Arch Dis Child 58:512–517
6. Vyas H, Field D, Milner AD, Hopkin IE 1986 Determinants of the first inspiratoryvolume and functional residual capacity at birth. Pediatr Pulmonol 2:189–193
7. Mortola JP, Fisher JT, Smith JB, Fox GS, Weeks S, Willis D 1982 Onset ofrespiration in infants delivered by cesarean section. J Appl Physiol 52:716–724
8. Saunders RA, Milner AD 1978 Pulmonary pressure/volume relationships during thelast phase of delivery and the first postnatal breaths in human subjects. J Pediatr93:667–673
9. Mortola JP, Fisher JT, Smith B, Fox G, Weeks S 1982 Dynamics of breathing ininfants. J Appl Physiol 52:1209–1215
10. Frappell PB, MacFarlane PM 2005 Development of mechanics and pulmonaryreflexes. Respir Physiol Neurobiol 149:143–154
11. Fisher JT, Mortola JP, Smith JB, Fox GS, Weeks S 1982 Respiration in newborns:development of the control of breathing. Am Rev Respir Dis 125:650–657
12. Kosch PC, Stark AR 1984 Dynamic maintenance of end-expiratory lung volume infull-term infants. J Appl Physiol 57:1126–1133
13. Lindroth M, Johnson B, Ahlstrom H, Svenningsen NW 1981 Pulmonary mechanicsin early infancy. Subclinical grunting in low-birth-weight infants. Pediatr Res15:979–984
14. Mortola JP, Milic-Emili J, Noworaj A, Smith B, Fox G, Weeks S 1984 Musclepressure and flow during expiration in infants. Am Rev Respir Dis 129:49–53
15. Radvanyi-Bouvet MF, Monset-Couchard M, Morel-Kahn F, Vicente G, Dreyfus-Brisac C 1982 Expiratory patterns during sleep in normal full-term and prematureneonates. Biol Neonate 41:74–84
16. Bolt RJ, van Weissenbruch MM, Lafeber HN, Delemarre-van de Waal HA 2001Glucocorticoids and lung development in the fetus and preterm infant. PediatrPulmonol 32:76–91
17. Jain L, Eaton DC 2006 Physiology of fetal lung fluid clearance and the effect oflabor. Semin Perinatol 30:34–43
18. Aly H, Massaro AN, Patel K, El Mohandes AA 2005 Is it safer to intubate prematureinfants in the delivery room? Pediatrics 115:1660–1665
19. Thomson MA 2002 Continuous positive airway pressure and surfactant; combineddata from animal experiments and clinical trials. Biol Neonate 81:16–19
20. Morley CJ, Davis PG, Doyle LW, Brion LP, Hascoet JM, Carlin JB 2008 NasalCPAP or intubation at birth for very preterm infants. N Engl J Med 358:700–708
21. Gerhardt T, Bancalari E 1980 Chestwall compliance in full-term and prematureinfants. Acta Paediatr Scand 69:359–364
22. Heldt GP, McIlroy MB 1987 Dynamics of chest wall in preterm infants. J ApplPhysiol 62:170–174
23. Barker PM, Gowen CW, Lawson EE, Knowles MR 1997 Decreased sodium ionabsorption across nasal epithelium of very premature infants with respiratory distresssyndrome. J Pediatr 130:373–377
24. Zelenina M, Zelenin S, Aperia A 2005 Water channels (aquaporins) and their role forpostnatal adaptation. Pediatr Res 57:47R–53R
25. Plakk P, Liik P, Kingisepp PH 1998 Hot-wire anemometer for spirography. MedBiol Eng Comput 36:17–21
26. Marsh MJ, Ingram D, Milner AD 1993 The effect of instrumental dead space onmeasurement of breathing pattern and pulmonary mechanics in the newborn. PediatrPulmonol 16:316–322
27. Morris MG 1999 A simple new technique to measure the effective dead space of theface mask with a water volumeter in infants. Eur Respir J 14:1163–1166
28. Morley C 2007 New Australian Neonatal Resuscitation Guidelines. J Paediatr ChildHealth 43:6–8
29. Stick S 1996 Measurements during tidal breathing. In: Stocks J, Sly PD, Tepper RS,Morgan WJ (eds) Infant Respiratory Function Testing. John Wiley & Sons Inc, NewYork, pp 118–119
30. Karlberg P, Cherry RB, Escardo FE, Koch G 1962 Pulmonary ventilation andmechanics of breathing in the first minutes of life, including the onset of respiration.Acta Paediatr 51:121–136
31. Engstrom L, Karlberg P, Rooth G, Tunell R 1966 The Onset of Respiration: A Studyof Respiration and Changes in Blood Gases and Acid-base Balance. Association forthe aid of crippled children, New York, pp 17–23
32. Lopes J, Muller NL, Bryan MH, Bryan AC 1981 Importance of inspiratory muscletone in maintenance of FRC in the newborn. J Appl Physiol 51:830–834
33. Stark AR, Cohlan BA, Waggener TB, Frantz ID III, Kosch PC 1987 Regulation ofend-expiratory lung volume during sleep in premature infants. J Appl Physiol62:1117–1123
34. Milner AD, Saunders RA, Hopkin IE 1978 Is air trapping important in the mainte-nance of the functional residual capacity in the hours after birth? Early Hum Dev2:97–105
35. Mortola JP, Magnante D, Saetta M 1985 Expiratory pattern of newborn mammals.J Appl Physiol 58:528–533
36. Long EC, Hull WE 1961 Respiratory volume-flow in the crying newborn infant.Pediatrics 27:373–377
37. Stocks J 1999 Lung function testing in infants. Pediatr Pulmonol Suppl 18:14–2038. Schmalisch G, Foitzik B, Wauer RR, Stocks J 2001 Effect of apparatus dead space
on breathing parameters in newborns: “flow-through” versus conventional tech-niques. Eur Respir J 17:108–114
39. Fleming PJ, Levine MR, Goncalves A 1982 Changes in respiratory pattern resultingfrom the use of a facemask to record respiration in newborn infants. Pediatr Res16:1031–1034
40. Dolfin T, Duffty P, Wilkes D, England S, Bryan H 1983 Effects of a face mask andpneumotachograph on breathing in sleeping infants. Am Rev Respir Dis 128:977–979
356 BREATHING PATTERNS AT BIRTH
A59
ARTICLE
Endotracheal Intubation Attempts During NeonatalResuscitation: Success Rates, Duration, and AdverseEffectsColm P.F. O’Donnell, MD, MRCPI, MRCPCHa,b, C. Omar F. Kamlin, MB, MRCP, MRCPCHa, Peter G. Davis, MD, FRACPa,b,Colin J. Morley, MD, FRACP, FRCPCHa,c
aDivision of Newborn Services, Royal Women’s Hospital, Melbourne, Victoria, Australia; bDepartment of Obstetrics and Gynecology, University of Melbourne, Melbourne,Victoria, Australia; cMurdoch Children’s Research Institute, Melbourne, Victoria, Australia
The authors have indicated they have no financial relationships relevant to this article to disclose.
ABSTRACT
OBJECTIVE.Endotracheal intubation of newborn infants is a mandatory competencefor many pediatric trainees. The Neonatal Resuscitation Program recommends a20-second limit for intubation attempts. Intubation attempts by junior doctors arefrequently unsuccessful, and many infants are intubated between 20 and 30seconds without apparent adverse effect. Little is known about the proficiency ofmore senior medical staff, the time taken to determine endotracheal tube (ETT)position, or the effects of attempted intubation on infants’ heart rate (HR) andoxygen saturation (SpO2) in the delivery room (DR). The objectives of this studywere to determine (1) the success rates and duration of intubation attempts duringDR resuscitation, (2) whether experience is associated with greater success ratesand shorter time taken to intubate, (3) the time taken to identify ETT position afterintubation, and (4) the frequency with which infants deteriorated during intuba-tion attempts and the time at which this occurred.
METHODS.We reviewed videos of DR resuscitations; identified whether intubationwas attempted; and, when attempted, whether intubation was attempted by aresident, a fellow, or a consultant. We defined the duration of an intubationattempt as the time from the introduction of the laryngoscope blade to the mouthto its removal, regardless of whether an ETT was introduced. We determined thetime from removal of the laryngoscope to the clinicians’ decision as to whether theintubation was successful and noted the basis on which this decision was made(clinical assessment, flow signals, or exhaled carbon dioxide [ETCO2] detection).We determined success according to clinical signs in all cases and used flow signalsthat were obtained during ventilation via the ETT or ETCO2 when available. Whenneither was available, the chest radiograph on admission to the NICU was re-viewed. For infants who were monitored with pulse oximetry, we determinedtheir HR and SpO2 before the intubation attempt. We then determined whethereither or both fell by !10% during the attempt and, if so, at what time it occurred.
www.pediatrics.org/cgi/doi/10.1542/peds.2005-0901
doi:10.1542/peds.2005-0901
KeyWordsinfant, newborn, resuscitation, positivepressure respiration, endotrachealintubation
AbbreviationsETT—endotracheal tubeNRP—Neonatal Resuscitation ProgramDR—delivery roomSpO2—oxygen saturationHR—heart rateETCO2—exhaled carbon dioxide
Accepted for publication Jul 20, 2005
Address correspondence to Colm P.F.O’Donnell, MD, MRCPI, MRCPCH, Division ofNewborn Services, Royal Women’s HospitalMelbourne, 132 Grattan St, Carlton, Victoria3053, Australia. E-mail: [email protected]
PEDIATRICS (ISSN 0031 4005). Copyright © 2006by the American Academy of Pediatrics
e16 O’DONNELL, et al
A60
RESULTS.We reviewed 122 video recordings in which oro-tracheal intubation was attempted 60 times in 31 in-fants. We secondarily verified ETT position using flowsignals, ETCO2, or chest radiographs after 94% of at-tempts in which an ETT was introduced. Thirty-seven(62%) attempts were successful. Success rates and mean(SD) time to intubate successfully by group were asfollows: residents: 24%, 49 seconds (13 seconds); fel-lows: 78%, 32 seconds (13 seconds); and consultants:86%, 25 seconds (17 seconds). Of the 23 unsuccessfulattempts, 13 were abandoned without an attempt to passan ETT and 10 were placed incorrectly. The time todetermine ETT position in the DR was longer whenclinical assessment alone was used. Infants who weremonitored with oximetry deteriorated during nearly halfof the intubation attempts. Deterioration seemed morelikely when HR and SpO2 were low before the attempt.
CONCLUSIONS. Intubation attempts often are unsuccessful,and successful attempts frequently take !30 seconds.Greater experience is associated with greater successrates and shorter duration of successful attempts. Flowsignals and ETCO2 may be useful in determining ETTposition more quickly than clinical assessment alone.Infants frequently deteriorate during intubation at-tempts. Improved monitoring of infants who are resus-citated in the DR is desirable.
INTERNATIONAL CONSENSUS STATEMENTS1 and guide-lines on neonatal resuscitation2,3 advise that infants
with inadequate respiration and/or bradycardia at birthbe given positive pressure ventilation with a manualventilation device with a face mask or endotracheal tube(ETT). Endotracheal intubation of newborn infants is amandatory competence for basic training in pediatrics inmany countries.4,5 The Neonatal Resuscitation Program(NRP) recommends that intubation attempts be limitedto 20 seconds.2
A United Kingdom study of the care of preterm in-fants and its effect on their survival identified difficultieswith intubation and the level of experience of staffpresent as the most common concerns about neonatalresuscitation.6 In a US study of pediatric residents, nonemet the authors’ definition of procedural competence forintubation (successful at first or second attempt !80%of the time) over a 2-year period.7 An additional USstudy of delivery room (DR) intubations that were per-formed mainly by residents and fellows found that fewwere successful within 20 seconds and that more infantswere intubated between 20 and 30 seconds withoutapparent adverse effect.8 Another study from the samecenter reported that just over half of all intubation at-tempts that were performed by junior doctors were suc-cessful and that residents there currently have inade-quate opportunity to become proficient.9
There is little information on the success rates orduration of intubation attempts of more experiencedoperators. Also, although adverse effects of endotrachealintubation in the intensive care setting have been de-scribed,10 there is little information on the effects ofintubation attempts on infants’ oxygen saturations(SpO2) and heart rate (HR) as determined by pulse oxim-etry in the DR. Using video recordings, we wished todetermine (1) the success rates and duration of intuba-tion attempts during neonatal resuscitation at our hos-pital, (2) whether the success rates and duration of suc-cessful attempts by staff varied with different levels ofexperience, (3) the time taken to identify ETT positionafter attempted intubation, and (4) the frequency withwhich infants deteriorated during intubation attemptsand the time at which this occurred.
METHODS
SettingThe Royal Women’s Hospital Melbourne is a tertiary-level perinatal center with "5000 deliveries per year.Approximately 450 infants are admitted to the NICU peryear, 100 to 110 of whom have a gestational age of #28weeks or birth weight of #1 kg. We do not have ad-vanced neonatal nurse practitioners or respiratory ther-apists; thus, all infants are intubated by medical staff:residents, fellows, or consultants. In general, the resi-dents are pediatric trainees whose first exposure to neo-natal medicine or intensive care occurs during their6-month rotation at our hospital. Usually, they have noprevious experience with intubation. The fellows havevariable but at least 18 months’ experience of neonatalintensive care and thus intubation. Consultants at ourhospital have a range of 5 to 35 years’ experience inneonatal medicine and extensive intubation experience.
Resuscitation PracticeAll staff who attend deliveries at our hospital completean in-house resuscitation training program based oninternational consensus statements1 and the NRP.2 TheNeopuff Infant Resuscitator (Fisher & Paykel, Auckland,New Zealand) t-piece and the Laerdal Infant Resuscitator(Laerdal, Oakleigh Victoria, Australia) self-inflating bagare the manual ventilation devices used at our hospital.Peak inflating and positive end-expiratory pressures areset at 30 and 5 cm H2O, respectively, on the Neopuff. Inthe DR, ETT position is judged from clinical signs asrecommended in international consensus statements1;on occasion, an exhaled carbon dioxide (ETCO2) detec-tor (Pedi-Cap Nellcor Puritan Bennet, Pleasanton, CA) isused for secondary confirmation. A time limit for intu-bation attempts is not enforced rigidly at our institution.Infants who are intubated in our DR are transferred tothe NICU, where they have a chest radiograph to con-firm ETT position.
PEDIATRICS Volume 117, Number 1, January 2006 e17
A61
Resuscitation StudiesSince January 2004, we have recorded DR resuscitationsat our hospital with the endorsement of our HumanResearch and Ethics Committee. When available, amember of the investigating team attended deliveries forwhich the need for resuscitation was anticipated andmade detailed recordings. When time allowed beforedelivery, the parents were approached and their permis-sion was sought to record the resuscitation. When timedid not allow, the resuscitation was recorded and theparents were approached as soon as practicable thereaf-ter. Their permission then was sought to view the re-cordings and to use them for data extraction and edu-cational purposes.
We recorded resuscitations using a digital video cam-era that was mounted above the resuscitation cot andpositioned to acquire a clear view of the infant andresuscitative interventions. Sound was audible fromthese recordings. The sensor of a Masimo Radical(Masimo Corp, Irvine, CA) pulse oximeter was placed onthe infants’ right hand or wrist as soon as practicableafter delivery. The oximeter was set to capture HR andSpO2 with maximal sensitivity and average data over2-second intervals.
Many of the infants who were videotaped had pul-monary mechanics measured during positive pressureventilation using the Florian Respiratory Monitor (Acu-tronic, Zug, Switzerland). This monitor measures pres-sures in the respiratory circuit directly and uses a sensorplaced between the ventilation device and the ETT tomeasure gas flow. These data were recorded and ana-lyzed using a laptop computer with Spectra software(Grove Medical, London, United Kingdom), a programdesigned specifically for the recording and analysis ofrespiratory signals (see Fig 1). Although not usuallyavailable to the resuscitation team in the DR, these sig-nals were used occasionally to determine ETT positionduring the resuscitation. They were used to determineETT position retrospectively for this study.
Evaluation of Video RecordingsAll video recordings were reviewed; those in which in-tubation was attempted were identified, and the reasonfor intubation was determined. The number of attempts,the timing of these attempts, and the grade of doctorwho performed the procedure were noted. In keepingwith previous reports,8 the duration of an intubationattempt was defined as the time from introduction of thelaryngoscope blade into the mouth to the time it wasremoved, irrespective of whether an ETT was intro-duced. When an ETT was introduced, we determined thetime from removal of the laryngoscope to the decision bythe resuscitating team whether the attempt was success-ful.
For infants who were monitored with pulse oximetryduring the procedure, we determined the times of life atwhich the sensor was applied and oximetry data wereavailable. We determined whether preoxygenation wasgiven before the attempt and the HR and SpO2 beforeand after the attempt. We considered the infants’ con-dition to have deteriorated when their HR and/or SpO2
fell by !10% during the procedure and noted the timeat which this occurred.
Statistical AnalysisMean (SD) duration of intubation attempts by differentgrades of doctors were compared using 1-way analysis ofvariance and test for linearity. Proportions of successfulintubations by different grades of doctors were com-pared using "2 test and linear-by-linear association. SPSSversion 12.0.1 (SPSS Inc, Chicago IL) was used for anal-ysis.
RESULTSWe attended 123 deliveries and were given parentalpermission to extract data from 122. Intubation wasattempted a total of 60 times in 31 infants whose mean(SD; range) gestational age and birth weight were 28weeks (5; 23–40) and 1227 g (939; 495–3870), respec-
FIGURE 1Pressure and flow traces recorded at 200 Hz using a flow sensor placed between the Neopuff and an ETT during positive pressure ventilation of an infant who was born at 26 weeks’gestation and weighed 621 g. Trace A was obtained after an esophageal intubation; no gas returns through the ETT when inflation stops. Trace B was obtained subsequently througha correctly placed ETT; gas returns through the ETT when inflation stops.
e18 O’DONNELL, et al
A62
tively. Twenty-one attempts were made by 12 residents,18 by 6 fellows, and 21 by 10 consultants. All wereorotracheal intubation attempts. The mean (SD; range)number of intubation attempts per infant was 2 (1; 1–5).The indications for intubation are shown in Table 1 andincluded treatment allocation in our randomized trialcomparing continuous positive airway pressure with en-dotracheal intubation and ventilation (the COIN trial).11
To be eligible for randomization, infants of 25 to 28weeks’ gestation had to have regular spontaneous respi-ration by 5 minutes of age. The infants for whom theindication for intubation was extreme immaturityand/or low birth weight per se had a mean (SD) gesta-tional age and birth weight of 24 weeks (1 weeks) and681 g (111 g), respectively.
In the DR, ETT position was determined by clinicalassessment alone after 46 (76%) attempts and by usingflow signals and ETCO2 on 7 (12%) occasions each. Weverified the ETT position retrospectively using flow sig-nals for 32 attempts and by examining the chest radio-graphs that were taken on admission to the NICU after 6attempts. We thus secondarily confirmed ETT positionusing flow signals, ETCO2, or chest radiograph after 94%(44 of 47) of attempts in which an ETT was introduced.
Successful AttemptsOverall, 37 (62%) intubation attempts were successful.In the DR, ETT position was determined by clinical as-sessment alone after 26 attempts; flow signals (see Fig1B) and ETCO2 were used in addition after 7 and 4attempts, respectively. Twenty-seven attempts weresubsequently verified as successful using flow signals.For the remaining 6 infants who had neither a flowsensor nor an ETCO2 detector in the circuit in the DR,review of the chest radiographs on admission to theNICU confirmed correct placement of the ETT.
The rates of success and duration of intubation at-tempts overall and by group are shown in Table 2. Thetime taken to intubate successfully differed betweengrades of doctors (P # .01, analysis of variance), withmore senior doctors intubating more rapidly (P # .01,test of linear trend; Table 2, Fig 3). Similarly, successrates differed between grades of doctors (P # .001, Pear-son "2), with senior doctors more successful (P # .001,linear-by-linear trend; Table 2). The 2 shortest successful
attempts were 8 seconds. Ten (17%) attempts were suc-cessful within 20 seconds, an additional 12 (20%) weresuccessful between 20 and 29 seconds, and the remain-ing 15 (25%) successful intubations took !30 seconds.The longest successful attempt took 70 seconds.
Of note, 1 successful intubation attempt was errone-ously thought not to be successful on clinical groundsand on cursory examination of the flow signals. Closeretrospective examination of these signals revealed thattiny quantities of gas were passing through the ETT (seeFig 2). In the DR, this correctly placed ETT was removedfrom this extremely preterm infant. The infant subse-quently was reintubated by an experienced fellow whowas confident that the ETT was positioned correctly. AnETCO2 detector was placed in the circuit but did notchange color. The ETT was confirmed to be passingthrough the cords on direct laryngoscopy by a consultant90 seconds after reintubation. The peak inflating pres-sure was increased from 30 to 40 cm H2O, and the HRincreased !100 beats per minute at 108 seconds. Theinfant’s own breathing effort improved at 131 seconds;only then was color change evident on the ETCO2 de-tector. This suggests that insufficient exhaled gas and,thus, CO2 passing through a correctly placed ETT was thereason that the reagent strip did not change color. This isthe second such “false-negative” result that we haveseen during the resuscitation of a preterm infant withnoncompliant lungs.12
FIGURE 2Traces recorded at 200 Hz during positive pressure ventilation that was given to an infantwhowas born at 28weeks andweighed 1354 g. The flow sensorwas placedbetween theETT andNeopuff, whichdelivered apeak inflatingpressure of 30 cmH2Oandpositive endexpiratory pressure of 5 cm H2O. Minute quantities (0.1 mL) of gas are delivered duringinflation and return when inflation stops.
TABLE 1 Indications for Intubation
Indication for Intubation No. of Infants(% of Total)(n $ 31)
No. of Attempts(% of Total)(n $ 60)
Bradycardia despite mask ventilation 9 (29) 16 (27)Extreme prematurity and/or low birth weight 10 (32) 27 (45)Congenital diaphragmatic hernia 3 (8) 3 (5)Treatment allocation in COIN trial 5 (16) 9 (15)Apnea/inadequate respiration 2 (7) 3 (5)Poor respiration and other congenital problem 2 (7) 2 (3)
TABLE 2 Success Rates and Duration of Attempts According to theGrade of Doctor Attempting Intubation
Total Residents Fellows Consultants
No. of attempts 60 21 18 21No. (%) of successful attempts 37 (62) 5 (24) 14 (78) 18 (86)Duration of attempts, mean(SD), s
33 (19) 38 (20) 36 (16) 28 (19)
Duration of successful attempts,mean (SD), s)
31 (17) 51 (13) 32 (13) 25 (17)
PEDIATRICS Volume 117, Number 1, January 2006 e19
A63
Unsuccessful AttemptsOf the 23 unsuccessful attempts, 13 were abandonedwithout an attempt to pass the ETT as the vocal cordscould not be visualized adequately. Seven attempts weredetermined to be unsuccessful by clinical assessmentalone in the DR. Of these, 4 were verified using flowsignals (see Fig 1A) and the infant could be heard to cryafter 2 attempts. ETCO2 was used in addition to clinicalassessment after 3 attempts. No intubation attempt wasdetermined unsuccessful using flow signals in the DR.
Time to Determination of ETT PositionOverall and excluding the 13 abandoned attempts forwhich no assessment of tube position was required, themean (SD; range) time taken to determine ETT positionwas 33 seconds (28; 1–127). The mean (SD [range]) timeto identify successful and unsuccessful attempts was notsignificantly different (32 [26] vs. 38 [37]; P $ .55). Themean (SD; range) time taken for clinical assessment ofETT position (n $ 33) was 39 seconds (30; 1–127); forflow signals (n $ 7) and ETCO2 (n $ 7), it was 19seconds (16; 8–53) and 17 seconds (20; 4–60), respec-tively. The time taken to assess tube position exceededthe time taken to intubate the infant on 15 (45%), 1(14%), and 2 (29%) occasions for which clinical assess-ment, flow signals, and ETCO2, respectively, were used.
Pulse Oximetry During Intubation AttemptsTwenty-seven infants had pulse oximetry during 51 in-tubation attempts. The oximetry sensor was applied tothe infant at a mean (SD) of 53(20) seconds of life, anddata were displayed at a mean (SD) of 80 (27) seconds oflife. The mean (SD; range) time at which intubation wasattempted was 253 seconds (135; 86–699).
Overall, infants deteriorated during 25 (49%) of these
attempts. Compared with the value just before the intu-bation attempt, SpO2 alone fell by !10% in 9, HR alonefell by !10% in 4, and both fell by !10% in 12. Thetime at which infants deteriorated during intubation wasvariable, ranging from 2 to 55 seconds (mean: 20; SD:13). Of the 12 attempts that were #20 seconds, no infantdeteriorated during the attempt. Infants deterioratedduring 4 of the 12 attempts that were between 20 and 29seconds and 20 of the 27 attempts that were !30 sec-onds. Infants were monitored during 30 successful intu-bation attempts and deteriorated during 14; during the21 failed attempts, 11 deteriorated.
As the reasons for intubation varied, so did the in-fants’ condition before the attempt, eg, infants who wereintubated for bradycardia despite mask ventilation hadlower HR than infants who were randomly assigned inthe COIN trial. Sixteen infants had HR #100 beforeintubation was attempted; 10 of these infants deterio-rated during the attempt. Of the 35 infants who had HR!100, 15 deteriorated. The mean SpO2 of the infants atthe time intubation was attempted was 70%. Of the 25infants with SpO2 #70%, 17 deteriorated; of the 26infants with SpO2 !70%, 8 deteriorated.
DISCUSSIONEndotracheal intubation of the newborn is an importantskill that seems to be difficult to acquire and improveswith experience. Routine intubation of infants who areborn through meconium-stained liquor is no longer rec-ommended; it is no longer acceptable to practice intu-bation on infants who have died; working hours fordoctors in training have been reduced internationally;and, in our hospital at least, there is a reduction inendotracheal ventilation as a result of increasing use ofcontinuous positive airway pressure. These factors meanthat the decline in opportunity to learn and practiceneonatal intubation and the consequent decline in pro-ficiency that has been described7 may well continue. Thishighlights the need for more anatomically suitable andlife-like mannequins of premature infants.
This study shows that intubation is successful only"60% of the time. Even when practitioners with con-siderable experience attempt intubation, many infantsare not intubated within the 20-second limit suggestedby the NRP.2 In addition to fellows who are alwayspresent, consultant neonatologists frequently attendhigh-risk deliveries at our hospital, reflected by the rel-atively high proportion (35%) of attempts made by thisgroup in our study. Our findings that consultants aremore likely to intubate infants successfully and morequickly support the practice of having experienced se-nior staff present for back-up at high-risk deliveries.
An interval that rarely is considered in the time takenfor intubation is the time taken for clinicians to decidewhether the ETT is placed correctly. In our study, thisfrequently took longer than intubation itself when clin-
FIGURE 3Box plot showing median, interquartile range, outliers, and extremes of duration of suc-cessful intubation attempts by grade of doctor.
e20 O’DONNELL, et al
A64
ical assessment alone was used. Our study suggests that,as has already been suggested for ETCO2,13,14 flow signalsmay be useful in determining ETT position more quicklythan clinical assessment alone. However, as also shownin this study, neither method is fool-proof, and bothrequire additional evaluation.
In our study, infants frequently deteriorated duringintubation attempts as determined by pulse oximetry. Itseems that the more unwell the infant at the time ofintubation, as indicated by the lower HR and SpO2, themore likely they were to deteriorate during the proce-dure. For most of these infants, deterioration was notreadily apparent on observation, and it is unlikely that itwould have been identified quickly by intermittent as-sessment of the HR, by either auscultation or palpation.Although infants deteriorated more frequently duringlonger intubation attempts, the timing of deteriorationwas highly variable. For example, of the 20 infants whodeteriorated during attempts of !30 seconds’ duration,12 had already done so by 20 seconds. Thus, rather thanimpose a general time limit for intubation attempts, weadvocate improved monitoring of infants in the DR (eg, theuse of pulse oximetry) and limiting the duration of intu-bation attempts according to their individual response.
CONCLUSIONSIntubation attempts are often unsuccessful, and success-ful attempts frequently require !30 seconds. Greaterexperience is associated with greater success rates andshorter duration of successful attempts. The time takento determine ETT position often exceeds the time takento intubate and may be reduced by assessing flow signalsor ETCO2. Infants frequently deteriorate during intuba-tion attempts. Improved monitoring of infants who areresuscitated in the DR is desirable.
ACKNOWLEDGMENTSDr O’Donnell received a Royal Women’s Hospital Post-graduate Degree Scholarship. Dr Davis is supported inpart by an Australian National Health and Medical Re-search Council Practitioner Fellowship.
REFERENCES1. International guidelines for neonatal resuscitation: an excerpt
from the guidelines 2000 for cardiopulmonary resuscitationand emergency cardiovascular care: International Consensuson Science. Pediatrics. 2000;106(3). Available at: www.pediatrics.org/cgi/content/full/106/3/e29
2. Kattwinkel J, ed. Textbook of Neonatal Resuscitation. 4th ed. ElkGrove, IL: American Academy of Pediatrics/American HeartAssociation Neonatal Resuscitation Program; 2000
3. Richmond S, ed. Resuscitation at Birth: Newborn Life SupportProvider Manual. London, United Kingdom: ResuscitationCouncil; 2001
4. Royal College of Paediatrics and Child Health UK. A frameworkof competences for basic specialist training in Paediatrics; 2004.Available at: www.rcpch.ac.uk/publications/education!and!training!documents/Competences.pdf. Accessed July 20,2005
5. Accreditation Council for Graduate Medical Education. Resi-dency Review Committees/Program Requirements/Pediatrics.Available at: www.acgme.org/acWebsite/downloads/RRC!progReq/320pr701.pdf. Accessed July 20, 2005
6. Confidential enquiry into stillbirths and deaths in infancy UK.Project 27/28: an enquiry into quality of care and its effect onthe survival of babies born at 27–28 weeks. Norwich, UnitedKingdom: Stationery Office; 2003. Available at: www.cemach.org.uk/publications/p2728/mainreport.pdf. Accessed July 20,2005
7. Falck AJ, Escobedo MB, Baillargeon JG, Villard LG, Gunkel JH.Proficiency of pediatric residents in performing neonatal endo-tracheal intubation. Pediatrics. 2003;112:1242–1247
8. Lane B, Finer NN, Rich W. Duration of intubation attemptsduring neonatal resuscitation. J Pediatr. 2004;145:67–70
9. Leone TA, Rich W, Finer NN. Neonatal intubation: success ofpediatric trainees. J Pediatr. 2005;146:638–641
10. Kelly MA, Finer NN. Nasotracheal intubation in the neonate:physiologic responses and effects of atropine and pancuro-nium. J Pediatr. 1984;105:303–309
11. Morley CJ, Davis PG, Doyle LW. Continuous positive airwayspressure: randomized controlled trial in Australia. Pediatrics.2001;108:1383
12. Kamlin COF, O’Donnell CPF, Davis PG, Morley CJ. Colorimet-ric end-tidal carbon dioxide detectors in the delivery room:strengths and limitations—case report. J Pediatr. 2005;147:547–548
13. Aziz HF, Martin JB, Moore JJ. The pediatric disposable end-tidal carbon dioxide detector role in endotracheal intubation innewborns. J Perinatol. 1999;19:110–113
14. Repetto JE, Donohue PK, Baker SF, Kelly L, Nogee LM. Use ofcapnography in the delivery room for assessment of endotra-cheal tube placement. J Perinatol. 2001;21:284–287
PEDIATRICS Volume 117, Number 1, January 2006 e21
A65
COLORIMETRIC END-TIDAL CARBON DIOXIDE DETECTORS IN THE
DELIVERY ROOM: STRENGTHS AND LIMITATIONS. A CASE REPORT
C. OMAR F. KAMLIN, MRCP, MRCPCH, COLM P. F. O’DONNELL, MRCPI, MRCPCH, PETER G. DAVIS, FRACP, MD,AND COLIN J. MORLEY, FRCP, FRCPCH, FRACP, MD
Clinical assessment and end-tidal carbon dioxide (ETCO2) detectors are used to verify tracheal intubation in newborn infants.
A case is presented in which an ETCO2 detector was misleading in determining endotracheal tube (ETT) position but useful in
determining the efficacy of ventilation in an extremely preterm infant. (J Pediatr 2005;147:547-8)
Extremely preterm infants are frequently intubated for ventilation. Determining whether the trachea has been intubated maybe difficult. End-tidal carbon dioxide (ETCO2) detectors, which indicate the presence of CO2 in exhaled gas by a colorchange, are useful for confirming correct endotracheal tube (ETT) placement in adults,1 children,2,3 and neonates.4 The
manufacturers of these detectors express caution about their use in infants weighing <1000 g and in low cardiac output states. Wedescribe a case in which a ETCO2 detector was misleading in determining ETT position but useful in determining the efficacy ofventilation in an extremely low birth weight infant.
CASE HISTORYA 32-year-old primigravida with a monochorionic, monoamniotic twin pregnancy presented at 26 weeks gestation.
Cardiotocography indicated that 1 twin had died and the other was severely compromised, prompting immediate delivery bycaesarean section under general anaesthetic. Twin boys were delivered. One infant was dead; the other was pale, apneic, andbradycardic and weighed 620 g. He was resuscitated by an experienced pediatrician, with the resuscitation videotaped. Maskventilation was started using the Neopuff infant resuscitator (Fisher & Paykel, Auckland, New Zealand), a flow-driven, pressure-limited T-piece that provides a consistent operator-selected peak inspiratory pressure (PIP) and positive end-expiratory pressure(PEEP). The flow rate was 8L/min of 100% oxygen, and the PIP and PEEP were initially set at 30 and 6 cm H2O, respectively.The infant was ventilated at 60/min. Prolonged inflations were not used. A Masimo Radical pulse oximeter (Masimo, Irvine,Calif ) was used to monitor preductal oxygen saturation (SpO2) and heart rate (HR). At 1 minute of age, the HR was 80 bpm, theSpO2 was 76%, and chest wall movement was not seen.
The infant was intubated with a 2.5-mm ETT by 2 minutes of age and ventilatedusing the same pressures. No chest wall movement was seen with inflations, the HRremained < 100 bpm, and SpO2 was 55%. The infant was then reintubated. His chest wallstill did not move, and his condition did not improve with ventilation, so an ETCO2
detector (Pedi-Cap; Nellcor Puritan, Bennett, Calif) was placed between the ETT andNeopuff to verify the ETT position. No color change was seen, suggesting that the ETTwas not in the trachea. At this time, HR was 75 bpm and SpO2 was 70%. The resuscitatorconfirmed correct placement of the ETT by direct laryngoscopy, then increased the PIPand PEEP in increments to 70 and 8 cm H2O, respectively. There was still no chest wallmovement with ventilation or color change in the Pedi-Cap.
At age 7 minutes, a 240-mL self-inflating bag (Laerdal, Wappingers Falls, NY) wasused with the pop-off valve occluded so that a very high (but unmeasured) PIP was used toinflate the lungs. Within 15 seconds, the Pedi-Cap detected sufficient exhaled CO2 tocause a color change. The HR then rose above 100 bpm, and chest wall movement was
ETCO2 End-tidal carbon dioxideETT Endotracheal tubeHR Heart rate
PEEP Positive end-expiratory pressurePIP Peak inspiratory pressureSpO2 Preductal oxygen saturation
547
From the Division of Newborn Ser-vices, Royal Women’s Hospital, andUniversity of Melbourne, Melbourne,Australia.Submitted for publication Feb 21,2005; last revision received Apr 19,2005; accepted May 6, 2005.Reprint requests: Dr. Omar Kamlin,Neonatal Services, Royal Women’sHospital, 132 Grattan Street, Carlton,VIC 3053, Australia. E-mail: [email protected]/$ - see front matterCopyrightª 2005 Elsevier Inc. All rightsreserved.
10.1016/j.jpeds.2005.05.014
A66
observed 30 seconds thereafter. Over the next 2 minutes, theSpO2 rose gradually to 85%. The PIP was reduced to produceshallow but obvious chest wall movement. The infant wastransferred to the neonatal intensive care unit receiving100% oxygen, with PIP and PEEP set at 45 and 8 cm H2O,respectively. No epinephrine, volume expansion, or externalcardiac massage was used during the resuscitation.
The infant’s initial hemoglobin was 2.8 g/dL, and adiagnosis of twin-to-twin transfusion syndrome was made.He was subsequently ventilated for 20 days without develop-ing air leaks and then treated with nasal continuous positiveairway pressure and a low fraction of inspired O2.
DISCUSSIONDelivery room intubation is difficult, with success rates of
<40% in infants <28 weeks gestational age.5 Signs of correctETT placement include direct visualization, chest wall move-ment with inflation, condensation inside the ETT, ausculta-tion of breath sounds, and improvement in HR and color.These observations are subjective, however, and sometimes canproduce false-positive impressions of tracheal ventilation.6
The pressures required to inflate the newborn lungare variable and unpredictable. They need to be above theairway critical opening pressure7 and may be >70 cm H2O.The International Liaison Committee on Resuscitation hasstated that visible chest wall expansion is a more reliable signof appropriate inflation pressures than in-line manometry onthe ventilation device.8 Others suggest that in very preterminfants, visible chest wall movement may represent excessivetidal volumes.5,9
The Pedi-Cap is a colorimetric semiquantitativeETCO2 detector used to verify correct ETT placement. Inthe presence of even small amounts of expired CO2, theindicator turns yellow, which reverses to purple in inspira-tion when the measured ETCO2 is <0.5% (4 mm Hg). Apersistent purple color is usually due to incorrect ETTplacement, but also may be seen in the setting of extremelylow pulmonary perfusion.10 In this case, the ETCO2 detectordid not change color because the infant had extremelynoncompliant lungs that could not be ventilated until veryhigh inflating pressures were used.
The ETCO2 detector gave the false impression that theETT was not in the trachea, which disagreed with theimpression gained from clinical assessment. Such false-negative results are a matter of concern, because they delay
appropriate intervention during resuscitation. Once theresuscitator was confident that the ETT was placed correctly,the inflating pressures were increased until ventilation wasachieved. The ETCO2 detector identified gas exchange beforethe chest wall was seen to move, suggesting that it is moresensitive for assessing adequacy of ventilation. Shortly afterexhaled CO2 was detected by the Pedi-Cap, the HR increased.This sequence suggests that the primary problem was inad-equate ventilation rather than low cardiac output leading topoor pulmonary perfusion.
Although evidence for the routine use of expired CO2
during neonatal resuscitation is limited, these devices arehighly sensitive in detecting tracheal intubation, and thereappears to be a low likelihood of causing harm.4,11 This caseillustrates that clinicians need to be aware that failure of thedevice to change color may be due to inadequate ventilationthrough a correctly placed ETT.
REFERENCES1. Goldberg JS, Rawle PR, Zehnder JL, Sladen RN. Colorimetricend-tidal carbon dioxide monitoring for tracheal intubation. Anesth Analg1990;70:191-4.2. Bhende MS, Thompson AE, Cook DR, Saville AL. Validity of adisposable end-tidal CO2 detector in verifying endotracheal tube placementin infants and children. Ann Emerg Med 1992;21:142-5.3. Bhende MS, Thompson AE. Evaluation of an end-tidal CO2 detectorduring pediatric cardiopulmonary resuscitation. Pediatrics 1995;95:395-9.4. Aziz HF, Martin JB, Moore JJ. The pediatric disposable end-tidalcarbon dioxide detector role in endotracheal intubation in newborns.J Perinatol 1999;19:110-3.5. Finer NN, RichWD. Neonatal resuscitation: raising the bar. Curr OpinPediatr 2004;16:157-62.6. Birmingham PK, Cheney FW, Ward RJ. Esophageal intubation:a review of detection techniques. Anesth Analg 1986;65:886-91.7. Upton CJ, Milner AD. Endotracheal resuscitation of neonates usinga rebreathing bag. Arch Dis Child 1991;66:39-42.8. Niermeyer S, Kattwinkel J, Van Reempts P, Nadkarni V, Phillips B,Zideman D, et al. International guidelines for neonatal resuscitation. Anexcerpt from the Guidelines 2000 for Cardiopulmonary Resuscitation andEmergency Cardiovascular Care: International Consensus on Science.Contributors and Reviewers for the Neonatal Resuscitation Guidelines.Pediatrics 2000;106:e29.9. Milner AD. Resuscitation at birth. Eur J Pediatr 1998;157:524-7.10. Roberts WA, Maniscalco WM, Cohen AR, Litman RS, Chhibber A.The use of capnography for recognition of esophageal intubation in theneonatal intensive care unit. Pediatr Pulmonol 1995;19:262-8.11. Repetto JE, Donohue P-CP, Baker SF, Kelly L, Nogee LM. Use ofcapnography in the delivery room for assessment of endotracheal tubeplacement. J Perinatol 2001;21:284-7.
548 Kamlin et al The Journal of Pediatrics ! October 2005
A67
Long versus short inspiratory times in neonates receivingmechanical ventilation (Review)
Kamlin COF, Davis PG
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library2009, Issue 1
http://www.thecochranelibrary.com
Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A68
T A B L E O F C O N T E N T S
1HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analysis 1.1. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 1 Mortality beforedischarge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Analysis 1.2. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 2 Air leak. . . . . . 18Analysis 1.3. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 3 BPD (supplemental oxygen
at 28 days). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Analysis 1.4. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 4 AaDO2 (mmHg) values
after 6 hours of intervention. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Analysis 1.5. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 5 IVH (all grades). . . 20Analysis 1.6. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 6 Patent ductus arteriosus
(PDA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Analysis 1.7. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 7 Cerebral palsy in survivors
less than 33 weeks gestation at birth. . . . . . . . . . . . . . . . . . . . . . . . . . . 21Analysis 1.8. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 8 Visual impairment in
survivors less than 33 weeks gestation at birth. . . . . . . . . . . . . . . . . . . . . . . 21Analysis 1.9. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 9 Sensorineural hearing loss
in survivors less than 33 weeks gestation at birth. . . . . . . . . . . . . . . . . . . . . . . 22Analysis 2.1. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds), Outcome 1
Mortality before discharge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Analysis 2.2. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds), Outcome 2 Air
leak. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Analysis 2.3. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds), Outcome 3
BPD (supplemental oxygenation at 28 days). . . . . . . . . . . . . . . . . . . . . . . . 23Analysis 2.4. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds), Outcome 4
IVH (all grades). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Analysis 3.1. Comparison 3 Long vs Short IT in HMD (subgroup analysis by diagnosis), Outcome 1 Mortality before
hospital discharge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Analysis 3.2. Comparison 3 Long vs Short IT in HMD (subgroup analysis by diagnosis), Outcome 2 Air leak. . . . 25Analysis 3.3. Comparison 3 Long vs Short IT in HMD (subgroup analysis by diagnosis), Outcome 3 BPD (supplemental
oxygen at 28 days). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Analysis 4.1. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation), Outcome 1
Mortality before discharge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26Analysis 4.2. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation), Outcome 2
Air leak. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Analysis 4.3. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation), Outcome 3
BPD (supplemental oxygen at 28 days). . . . . . . . . . . . . . . . . . . . . . . . . . 28Analysis 4.4. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation), Outcome 4
IVH (all grades). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2929WHAT’S NEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
iLong versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A69
29CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
iiLong versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A70
[Intervention Review]
Long versus short inspiratory times in neonates receivingmechanical ventilation
Camille Omar Farouk Kamlin1, Peter G Davis2
1Department of Neonatology, Royal Women’s Hospital, Victoria, Australia. 2Department of Paediatrics, Royal Women’s Hospital,Parkville, Australia
Contact address: Camille Omar Farouk Kamlin, Department of Neonatology, Royal Women’s Hospital, Gratton Street, Carlton,Victoria, 3053, Australia. [email protected].
Editorial group: Cochrane Neonatal Group.Publication status and date: Edited (no change to conclusions), published in Issue 1, 2009.Review content assessed as up-to-date: 22 June 2003.
Citation: Kamlin COF, Davis PG. Long versus short inspiratory times in neonates receiving mechanical ventilation. Cochrane Databaseof Systematic Reviews 2003, Issue 4. Art. No.: CD004503. DOI: 10.1002/14651858.CD004503.pub2.
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A B S T R A C T
Background
When intermittent positive pressure ventilation (IPPV) was introduced in newborn infants with hypoxic respiratory failure from hyalinemembrane disease (HMD), mortality was high and air leaks problematic. This barotrauma was caused by the high peak inspiratorypressures (PIP) required to oxygenate stiff lungs. The primary determinants of mean airway pressure (and thus oxygenation) on aconventional ventilator are the inspiratory time (IT), PIP, positive end expiratory pressure and gas flow rates. In the 1970s uncontrolledstudies on a small number of infants demonstrated a benefit in reducing barotrauma using a long IT and slow rates. This strategy wassubsequently widely adopted. Current neonatal ventilators have been designed to minimise lung injury but rates of bronchopulmonarydysplasia (BPD) remain high. It is therefore important that the inspiratory time causing least harm is used.
Objectives
To determine in mechanically ventilated newborn infants whether the use of a long rather than a short IT reduces the rates of death,air leak and BPD.
Search strategy
The standard search strategy of the Cochrane Neonatal Review Group (CNRG) was used. Searches of electronic and other databaseswere performed. These included MEDLINE (1966 - April 2004) and the Cochrane Central Register of Controlled Trials (CENTRAL,The Cochrane Library, Issue 4, 2003). In order to detect trials that may not have been published, the abstracts of the Society forPediatric Research, and the European Society for Pediatric Research were searched from 1998 - 2003.
Selection criteria
All randomised and quasi-randomised controlled trials enrolling mechanically ventilated infants with or without respiratory pathologyevaluating the use of long versus short IT (including randomised crossover studies with outcomes restricted to differences in oxygenation).
Data collection and analysis
The standard method of the Cochrane Collaboration and its Neonatal Review Group were used. Two authors independently assessedeligibility, and the methodological quality of each trial, and extracted the data. The data were analysed using relative risk (RR) and riskdifference (RD) and their 95% confidence intervals. A fixed effect model was used for meta-analyses.
1Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A71
Main results
In five studies, recruiting a total of 694 infants, a long IT was associated with a significant increase in air leak [typical RR 1.56 (1.25,1.94), RD 0.13 (0.07, 0.20), NNT 8 (5, 14)]. There was no significant difference in the incidence of BPD. Long IT was associatedwith an increase in mortality before hospital discharge that reached borderline statistical significance [typical RR 1.26 (1.00, 1.59), RD0.07 (0.00, 0.13)].
Authors’ conclusions
Caution should be exercised in applying these results to modern neonatal intensive care, because the studies included in this reviewwere conducted prior to the introduction of antenatal steroids, post natal surfactant and the use of synchronised modes of ventilatorysupport. Most of the participants had single pathology (HMD) and no studies examined the effects of IT on newborns ventilated forother reasons such as meconium aspiration and congenital heart disease (lungs with normal compliance). However, the increased ratesof air leaks and deaths using long ITs are clinically important; thus, infants with poorly compliant lungs should be ventilated with ashort IT.
P L A I N L A N G U A G E S U M M A R Y
Long versus short inspiratory times in neonates receiving mechanical ventilation
Endotracheal intubation and positive pressure ventilation of newborn infants with respiratory failure has revolutionised neonatalintensive care. The majority of infants are ventilated because of lung immaturity and hyaline membrane disease, respiratory difficultiesthat resolve for most of these infants. Use of ventilators can cause lung inflammation and ventilator-induced lung injury, particularlyin the neonate with compliant chest walls, highlighting the importance of protective ventilation strategies. In addition, an infant’sown breathing efforts, when combined with ventilator inflations, can exacerbate lung injury. Managing the mean airway pressure isimportant to improve oxygenation and not cause excessive airway pressures that can damage the lungs and impede venous return (afactor implied in spontaneous intraventricular haemorrhage in preterm infants). Clinicians still need to set an inspiratory time on theventilator, making it important to determine whether the use of a long rather than a short inspiratory time reduces the rates of death,air leak and bronchopulmonary dysplasia (BPD, requiring supplemental oxygen at 28 days) in mechanically ventilated newborn infants(term and preterm). The review authors identified five randomised studies reported from 1980 to 1992. These trials recruited a total of694 newborn infants with acute respiratory failure mainly caused by hyaline membrane disease. A long inspiratory time was associatedwith a significant increase in air leak from the lungs (NNT 8). There was no significant difference in the incidence of BPD but anincrease in mortality before hospital discharge reached borderline statistical significance. Caution should be exercised in applying theseresults to modern neonatal intensive care because these studies were conducted before the introduction of antenatal steroids, postnatalsurfactant and the use of synchronised modes of ventilatory support. Whilst there is increasing use of non-invasive ventilation such asnasal continuous positive airway pressure to avoid ventilator-induced lung injury, mechanical ventilation will continue to have a rolein extremely immature infants and those with hyaline membrane disease complicated by apnea.
B A C K G R O U N D
The application of techniques of endotracheal intubation and in-termittent positive pressure ventilation (IPPV) for newborns withrespiratory failure revolutionised neonatal intensive care. Prior tothe introduction of IPPV, the mainstay of early management wassupportive care, including the delivery of supplemental oxygen,maintenance of normoglycaemia and correction of metabolic aci-dosis. These measures had been recognised to reduce morbiditycompared to controls (Usher 1963). IPPV was introduced as res-cue therapy to treat apnea or correct hypoxaemia and respiratoryacidosis in infants likely to die, often after bag and mask venti-
lation had failed. The techniques applied were initially adaptedfrom those used in adult ventilation and met with limited successwhen first introduced (D-Papadopoulos 1965, Murdock 1970).The short inspiratory times used in adults to avoid impaired ve-nous return and the fast rates to avoid asynchrony in newbornswere not as effective in preterm infants with atelectatic and poorlycompliant lungs (Adamson 1968). Gregory 1971 observed dra-matic improvements over earlier approaches in neonates with res-piratory distress syndrome (RDS) when continuous positive air-way pressure (CPAP) was applied through an endotracheal tube.Ventilators were introduced in the early 1970s which used cir-
2Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A72
cuits with continuous gas flow and a timing device to close theexpiratory valve allowing the infant to breathe spontaneously andreceive intermittent mandatory ventilation (IMV). Cumarasamy1973 reported improved survival when IMV was combined withthe application of positive end expiratory pressure (PEEP), al-though early experiences in this era document the use of high peakinspiratory pressures (PIP) to achieve adequate oxygenation. Thehigh PIPs which were used reflected the poor compliance of thelung and were associated with high rates of mortality and bron-chopulmonary dysplasia (BPD) (Northway 1967).
Alternative strategies were sought to treat infants with HMD us-ing lower inflating pressures. Reynolds 1971 and Herman 1973noted that alveolar recruitment and hence oxygenation could beimproved by using low rates (30 - 40 breaths per minute) and longinspiratory to expiratory (I:E) ratios. In some infants, improve-ments in oxygenation were seen when inspiratory times exceededthe expiratory times (inversed I:E ratios). In inverse- ratio ventila-tion the IT is prolonged, thereby increasing mean airway pressuresand allowing the use of lower PIP. As neonatal ventilation becamemore widely used in nurseries in the late 1970s/early 1980s, thisstrategy was often adopted even though the evidence of efficacywas based on small numbers (less than ten in both studies) withonly improved arterial blood gases and indices of oxygenation re-ported as the primary outcomes.
In assisted ventilation, the ventilator maintains oxygenation byproviding an effective mean airway pressure (MAP) exceeding at-mospheric pressure and a suitable inspired oxygen concentration.It removes carbon dioxide by passively removing the delivered tidalvolume. The tidal volume is the volume of gas moving in and outof the lung in a respiratory cycle. The minute ventilation is theproduct of tidal volume and respiratory frequency. The alveolarventilation (minute ventilation minus dead space ventilation) de-termines the amount of carbon dioxide removed. Ventilator “cy-cling” refers to the change from inspiration to expiration and maybe initiated by time (a pre-determined duration of inspiration),pressure (when a preset pressure is reached) or volume (expirationstarts after a preset tidal volume is delivered).
Early pressure limited time cycled (PLTC) ventilators produced asquare wave when pressure (on the y axis) was plotted against time(on the x axis). MAP is equal to the integration of the area un-der the pressure-time curve during a single respiratory cycle. Thesquare wave is dependent on the pre-set inspiratory time, gas flowrate and ventilator frequency. Reynolds 1971 used square waveanalysis to hypothesise that prolonging inspiratory time and usingslower rates would improve oxygenation by increasing MAP. Thelonger the IT, the longer the lungs are held distended at the setPIP. Sustained inspiratory pressures may have their greatest incre-mental effect on non-functioning lung units, thereby improvingdistribution of the delivered tidal volume. With high frequency
positive pressure ventilation (HFPPV; rates >60) changes in thepressure waveform are seen with a shortening of the inspiratoryplateau such that the square wave is replaced by a sine wave. Thusany change in IT that accompanies frequency adjustments willchange pressure waveforms, thereby altering MAP and oxygena-tion. Increasing MAP has been shown to improve oxygenation inRDS (Boros 1979). In contrast to early work by Reynolds 1971,Stewart 1981 found that increasing PEEP had the greatest effecton MAP. Whilst a high MAP may be useful in acute RDS, exces-sive airway pressures in the recovering lung may impede venousreturn and cause overdistension.
The time required for the lungs to inflate and deflate is determinedby the compliance and resistance of the respiratory system whichincludes the ventilator circuit, endotracheal tube and the patient’slung. The product of resistance and compliance, the time constant,is a measure of how long it takes for alveolar and proximal air-way pressures to equilibrate. It provides a theoretical basis on howbest to divide the respiratory cycle into inspiratory and expiratorytimes. The inspiratory time constant is typically short (approxi-mately 0.05 seconds) in RDS and relatively long (0.25 seconds) ininfants with normal lungs. For practical purposes, an expiratorytime equivalent to three time constants must be provided to allow95% of inspired tidal volume to be expelled (Harris 1996). Thusif the expiratory times are absolutely or relatively short (as seen ininversed I:E ratios) there is the potential for gas trapping to occurand the build up of pressure known as inadvertent PEEP (Weigl1973). Potential complications of inadvertent PEEP include gastrapping, reduced compliance and air leak. This can lead to re-duced pulmonary blood flow and central venous return, a factorimplied in spontaneous intraventricular haemorrhage in preterminfants (Rennie 1987). Ahluwalia 1994 measured spontaneous in-spiratory and expiratory times in newborns ventilated for RDS ata median of 0.3 (range 0.26- 0.34) and 0.46 (0.34-0.66) secondsrespectively. RDS, characterised by markedly reduced compliancewith very short time constants, can theoretically be managed byusing either rapid rates (>60) or slow rates with long inspiratorytimes providing there is sufficient time for passive exhalation. WithHFPPV, the expiratory time may be insufficient to allow for defla-tion and lung units can become overdistended and further reducethe compliance of the lung.
Meconium aspiration affects mature infants and is characterisedby both parenchymal and airway disease. Inhalation of meconiumor congenital pneumonias can present as non homogenous lungdisease typified by adjacent areas of atelectasis and hyperinflation.Setting ventilator parameters for non homogenous lung diseasewith different time constants is challenging with the aim beingto provide sufficient alveolar ventilation without the developmentof inadvertent PEEP. The measurement of compliance and resis-tance in these circumstances is then dependent on the frequencyof ventilation. With the limitations of PLTC ventilators, it hasbeen observed that tidal volumes decrease as inspiratory times are
3Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A73
reduced beyond a critical point (less than the time constant ofthe respiratory system) and that above a certain rate (>75), deadspace ventilation increases and minute ventilation is not a linearfunction of frequency (Boros 1984).
Recent advances in neonatal intensive care may influence the op-timal inspiratory time. These include the availability of exogenoussurfactant therapy (Soll 2004, Yost 2003) and the advances in ven-tilator technology that allow synchronisation of ventilator breathswith an infant’s spontaneous breathing (Greenough 2004). Sur-factant improves oxygenation by increasing functional residual ca-pacity. By stabilising patent airways and with the gradual recruit-ment of atelectatic regions of the lung, surfactant improves respi-ratory system compliance and may alter the optimal inspiratorytime (Goldsmith 1996). Attempts to achieve synchrony have in-cluded the use of muscle relaxation (Greenough 1986) and rapidrates to avoid active expiration against positive pressure inflation (Greenough 2004). Modern neonatal ventilators achieve synchro-nisation by assisting each spontaneous breath or providing syn-chronised intermittent mandatory breaths. In addition, new gen-eration neonatal ventilators with the addition of flow sensors allowcontinuous respiratory function and lung volume measurements.By measuring lung volumes, tidal volumes, resistance, dynamicand static lung compliance, much of the guesswork of setting ven-tilator parameters has been eliminated. However, despite these ad-vances in ventilator technology, clinicians still need to set an in-spiratory time.
The many variables involved in conventional mechanical ventila-tion (PIP, PEEP, IT, rate, flow rates) are interdependent making itdifficult to assess the effects of changing one variable (IT) whilstholding the others constant. By using stated subgroup analyses(see objectives), the aim of this review is to identify the optimalinspiratory time (or range of times) in newborn infants needingrespiratory support on conventional mechanical ventilation.
O B J E C T I V E S
The primary objective was to evaluate whether, in mechanicallyventilated neonates, the use of long inspiratory times comparedwith short inspiratory times improved short and long term out-comes. These include mortality and the rate of acute lung injury(air leak) and/or chronic lung injury (BPD). For the initial analy-sis, any definitions of short and long used by the authors of studieswere used.
Planned sub-group analyses included:
1. A subgroup of studies defining short IT as less than or equal to0.5 seconds
2. Subgroup analysis based on the overall ventilator strategy i.e.
a. Absolute IT set at different (long and short) levels and main-tained constant through the infant’s treatment
b. I:E ratios set at different levels and maintained constant so thatwhen rate was altered the IT and expiratory time were adjusted tomaintain the I:E ratio
c. Trials where IT or I:E ratio form only part of the ventilatorstrategy e.g. when absolute IT and I:E ratios were not prescribed.In these trials an acceptable range of ITs and I:E ratios as definedby the authors would have been used in each arm
3. Underlying pathology (RDS vs other causes)
4. Gestational age i.e. term vs preterm
5. Use of muscle relaxants vs none
6. Surfactant vs no surfactant
7. Mode of ventilation (synchronised vs non synchronised)
M E T H O D S
Criteria for considering studies for this review
Types of studies
All randomised and quasi-randomised controlled trials. Ran-domised crossover studies were eligible, but only for the assess-ment of acute effects on oxygenation.
Types of participants
Term and preterm infants less than 28 days of age and requiringconventional mechanical ventilation. No restrictions on underly-ing pathophysiology were applied.
Types of interventions
Short versus long inspiratory times in conventional mechanicalventilators including time cycled pressure limited ventilators (syn-chronised and non synchronised) and volume cycled ventilation.Trials using high frequency ventilation (> 150 breaths per minute)were excluded.
Types of outcome measures
PrimaryMortality in the first month of life, and before hospital dischargeIncidence of acute lung injury such as rates of all air leaks(pneumothorax, pulmonary interstitial emphysema, pneumome-diastinum, pneumoperitoneum)Incidence of chronic lung injury (rates of BPD)
• The need for supplemental oxygen at 28 days of life
4Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A74
• The need for supplemental oxygen at 36 weekspostmenstrual age
SecondaryIndices of improved oxygenation including
• Oxygenation index (OI)• Alveolar- arterial oxygen difference (AaDO2)• PaO2/FiO2 ratio (P/F ratio)
Lung mechanics (compliance and resistance of either the lung orrespiratory system)Duration of positive pressure respiratory support in daysDuration of supplemental oxygen requirement in daysRadiological evidence of neurologic injury
• IVH (All grades and severe grades 3 or 4 intraventricularhaemorrhage by Papile 1978 classification)
• Periventricular leukomalacia (PVL)
Long term neurodevelopmental outcomes in childhood (blind-ness, sensorineural deafness, developmental delay on recognisedpsychometric testing)
Search methods for identification of studies
See: Cochrane Neonatal Research Group (CNRG) search strategyThe standard search strategy of the CNRG was used. These in-cluded MEDLINE (1966 - April 2004) and the Cochrane CentralRegister of Controlled Trials (CENTRAL, The Cochrane Library,Issue 4, 2003). The MEDLINE search terms used were newborn,neonate, infant, mechanical ventilation, positive pressure respira-tion and the text phrase “ventilator rate” and text words “inspira-tory” or “expiratory”. No language restrictions were applied. Theabstracts of the Society for Pediatric Research, and the EuropeanSociety for Pediatric Research were searched from 1998 - 2003.
Data collection and analysis
The standard methods of the Cochrane Collaboration and theCNRG were used. The methodological quality of each trial wasreviewed independently by the two authors for blinding of ran-domisation, blinding of intervention and outcome measurementsas well as completeness of follow up.The two authors independently assessed eligibility of retrievedstudies and extracted data; the results were compared and differ-ences resolved by discussion.The statistical analysis used the fixed effect model. For categori-cal data the relative risk (RR), risk difference (RD) and numberneeded to treat (NNT) with 95% confidence intervals were cal-culated. Continuous data were analysed using weighted mean dif-ference (WMD). Where data for individual outcomes was incom-plete (i.e. did not include all randomised babies), numerators anddenominators were those presented in the published manuscript.
R E S U L T S
Description of studies
See: Characteristics of included studies; Characteristics of excludedstudies.For the summary of the included studies see also table of includedstudies.Five published trials (Spahr 1980, Heicher 1981, Greenough1981, OCTAVE 1991, Pohlandt 1992) recruiting a total of 694infants were identified. All trials were performed in the pre-surfac-tant era and when antenatal steroids were not widely used. Three (Spahr 1980, Heicher 1981 and Greenough 1989) were single cen-tre studies whilst OCTAVE 1991 and Pohlandt 1992 involved 6and 7 centres respectively.One study (Nilmeier 1995) was excluded. This was a randomisedcontrolled trial in infants with ventilator dependent HMD re-cruited at two weeks of age, comparing two ITs of 1.0 and 0.4seconds over a period of seven days. Oxygenation and respiratorymechanics were measured prior to randomisation and comparedafter a week of the intervention. The results, however were onlypublished in part and were descriptive in nature. Improvementsin oxygenation were described as indicated by FiO2 requirementrather than our prespecified indices (AaDO2, PF ratios), and thesewere only presented for infants randomized to a long IT.Study PopulationThe inclusion criteria varied slightly between studies. All the stud-ies except Greenough 1989 recruited infants with acute respiratoryfailure. Greenough 1989 compared different ITs when weaninginfants recovering from HMD. Only Pohlandt 1992 had specificgestational age criteria for inclusion, recruiting infants less thanor equal to 32 weeks gestation at birth. Heicher 1981 recruitedventilated infants with a birth weight of greater than 750g. In thestudy of OCTAVE all ventilated infants less than 72 hours of agewere eligible whilst Spahr 1980 recruited all ventilated newbornsover a 12 month period. The age at recruitment was not stated inthe studies of Heicher 1981 and Spahr 1980. The differences inunderlying respiratory pathology between the studies were small.Most infants had HMD although some cases of pneumonia wereincluded. Only OCTAVE 1991 and Heicher 1981 specifically ex-cluded infants with meconium aspiration syndrome (MAS) butno data specific to MAS were provided by any of the includedstudies.InterventionsThe ventilator strategies and types of ventilator used (see table ofincluded studies) varied with each study.Once randomised to a long IT or short IT, all infants were venti-lated to meet therapeutic range of oxygenation and eucapnia as de-fined by the authors. All studies allowed adjustments in PIP to cor-rect hypoxia. Ventilation was weaned by decreasing PIP. Where theIT was held constant, weaning of the ventilator rate was achieved
5Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A75
by lengthening the expiratory time. All studies had a priori failurecriteria (see table).Inspiratory times as defined by the authors were as follows;
• Spahr 1980 2.0 vs 1.0 seconds. Both groups were ventilatedat a rate of 20 breaths per minute (bpm). This allowed acomparison of I:E ratios (2:1 vs 1:2). IT ratios were held constantthroughout whilst rates were allowed to be increased to 40 bpmto correct acidosis and hypercarbia. The range of ITs compared inthis study were thus 2.0 - 1.0 (long) vs 1.0 - 0.5 (short) seconds.
• Heicher 1981 1.0 vs 0.5 seconds. The ITs were heldconstant throughout the study period. The long IT group wereventilated at a rate of 20 - 40bpm and an IT of 1.0 seconds. Therate was increased to correct hypercarbia by reducing expiratorytime. The short IT group were ventilated at a rate of 60 bpm andan IT of 0.33 seconds. Both groups were weaned by decreasingPIP and rate by increasing the expiratory time only. I:E ratioswere thus not held constant.
• OCTAVE 1991 1.0 - 0.75 vs 0.33 seconds. For the long ITgroup, starting rate was 30 (range 20-40) bpm using an I:E of1:1 and IT of 1.0 seconds. Whilst weaning the IT was reduced to0.75 seconds. Thus in the long IT group neither IT or I:E ratiowere held constant. The short IT group had an IT of 0.33seconds, starting rate of 60 bpm, fixing the I:E at 1:2. The ITand I:E ratio was constant throughout in this arm inuncomplicated cases. In the presence of severe hypoxia as definedby the authors, IT could have been increased to 0.5 seconds withthe rate unchanged giving an I:E of 1:1 in the short IT group.
• Pohlandt 1992 1.0 - 0.66 vs 0.33 seconds. For the long ITgroup, the starting rate was 30 bpm, an IT of 1 second with I:Eof 1:1. The IT was reduced to 0.66 seconds when weaning andrate decreased by increasing expiratory time. Thus in the long ITgroup neither IT or I:E ratio were held constant. Infants in theshort IT group were ventilated with an IT of 0.3 seconds at arate of 60bpm, fixing the I:E at 1:2. Infants in this group wereweaned by reducing PIP and PEEP until the infant was receivingendotracheal continuous positive airway pressure (CPAP).
• Greenough 1989 0.7 -1.0 vs 0.5 seconds. Infants withHMD were recruited if treated with HFPPV and rate had beenreduced to 60 bpm. IT for both groups was 0.5 seconds atrecruitment. The long IT group were weaned (reduction ofrespiratory rate) by adjusting both IT and expiratory time andkeeping I:E constant at 1:1.2. The IT was increasedincrementally to 1.0 seconds. In the short IT group IT was heldconstant at 0.5 seconds and I:E was variable. Rate was graduallyreduced to zero (endotracheal CPAP). A crossover to the otherarm was permitted for 30 minutes when rates were weaned to 40and 20 bpm, allowing a comparison of arterial blood gases at agiven rate and I:E ratio using long and short ITs.
Major Outcomes The major and secondary outcomes as stated bythe authors were as follows.
Spahr 1980 and Heicher 1981 - mortality and morbidity (air leakand BPD) and the effect of two different ITs on ventilator settings(MAP, PIP, PEEP and rate compared to baseline). OCTAVE 1991- mortality and morbidity (air leak and BPD). This study alsofollowed up surviving premature infants (<33 weeks) recruited toassess the effect of treatment allocation on long term neurodevel-opmental outcomes. Pohlandt 1992- the incidence of air leak fol-lowing randomisation. Mortality and the incidence of BPD wererecorded as secondary outcomes. In the study of Greenough 1989the primary outcome was duration of weaning. Secondary out-comes included rates of air leak, death and the need for re-venti-lation.
Risk of bias in included studies
The methodological quality of each trial was assessed using thecriteria of the Neonatal Review Group. For each trial the followingwere assessed: concealment of allocation schedule, inclusion in theanalysis of all randomised participants and blinding of interven-tion and outcome measurement.1. Concealment of allocation scheduleOCTAVE 1991 was a multi-centre trial in which the randomi-sation was adequately concealed by the use of sealed opaque en-velopes held at the co-coordinating centre. Participating centresmade a telephone call to the co-ordinating centre when an infantwas enrolled and the next sealed envelope opened. Methods ofconcealing the allocation schedule were not used in the studies ofHeicher 1981 and Spahr 1980. Lists were constructed in the studyof Pohlandt 1992 such that the numbers treated with either a longor short IT were equal after every second infant enrolled for a par-ticular gestation. These lists were not concealed. The method usedin randomising enrolled infants was not described by Greenough1989 and thus concealment of allocation was unknown.Generation of allocation sequence -This was not stated for OCTAVE 1991. Pohlandt 1992 used ran-domisation tables divided into blocks of two for each interven-tion arm for each of the intended target population (under 28weeks, 28-30 and 31-32 weeks). Spahr 1980 used a coin toss todetermine the treatment allocation sequence. Heicher 1981 quasi-randomised enrolled infants using alternate selection for long andshort ITs. The generation of allocation sequence was not stated inthe study of Greenough 1989.
2. Inclusion in the analysis of all randomised participantsAll participants were accounted for in the included studies.Pohlandt 1992 used an unusual recruitment and randomisationprocedure. Consent was not obtained and all preterm infants lessthan 32 weeks needing mechanical ventilation were provisionallyenrolled and randomised. Infants not meeting secondary enrol-ment criteria (FiO2 > 0.4) were excluded and replaced by the nextpreterm infant in the same gestational age category (37 of 181 ini-tially randomised). An additional seven infants were excluded after
6Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A76
randomisation because they received the alternate form of therapyduring transport. Spahr 1980 excluded five (of 74) patients postrandomisation either because of air leaks (one before randomisa-tion, two post PDA ligation), one because of sepsis and one be-cause of use of a muscle relaxant. The remaining three studieshad no post randomisation exclusions. OCTAVE 1991 followeda subgroup of enrolled patients (<33 weeks gestation at birth) ata median age of 18 months. Of the 183 surviving eligible infants,175 (96%) were seen.3. Blinding of interventionIt was not possible to blind the caregivers in any of the studies.4. Blinding of outcome assessmentsAll the primary outcome assessments were unblinded in the five in-cluded studies with one exception (Pohlandt 1992), where physi-cians and radiologists blinded to treatment allocation diagnosedair leak and BPD. In the OCTAVE 1991 study 98% of the longterm neurodevelopmental assessments were blinded.
Effects of interventions
Five studies (Spahr 1980, Heicher 1981, Greenough 1989,OCTAVE 1991, Pohlandt 1992) enrolling 694 infants met theentry criteria and were included in the analysis.Primary Outcomes
Comparison 01: Long versus short IT as defined by investigators(all trials)
MortalityMortality before hospital discharge was reported in all includedstudies (Table 1). OCTAVE 1991 was the only study to reportmortality after hospital discharge. In this study mortality at latestfollow up was 30/126 (24%) in the long IT group comparedto 36/125 (29%) in the short IT group. In a pooled analysis,there was a trend toward an increased rate of mortality beforehospital discharge in newborns ventilated with a long IT thatreached borderline statistical significance [typical RR 1.26 (1.00,1.59), typical RD 0.07 (0.00, 0.13)].Air LeakThe rates of air leak (acute lung injury) were reported in all in-cluded studies (table 2). The studies of Heicher 1981 and Pohlandt1992 demonstrated statistically significant differences in this out-come disfavouring the use of a long IT. There were no signifi-cant differences in the studies of Greenough 1989, Spahr 1980.OCTAVE 1991, the largest individual study in this meta-analysis,reported no overall difference in the rates of air leak but when thisoutcome was compared within a subgroup of infants with HMD,there was a statistically significant difference favouring the use of ashort IT (p= 0.013). In a pooled analysis, the use of a long IT wasassociated with a significantly increased rate of air leak [typical RR1.56 (1.25, 1.94), typical RD 0.13 (0.07, 0.20), NNT 8 (5, 14)].BPD (Chronic lung injury)
The rates of BPD in all included studies (Table 3) was defined as theneed for supplemental oxygen at 28 post natal days. None of thestudies reported on this outcome at 36 weeks post conceptual age.No individual study found a significant difference in this outcomeand in a pooled analysis, there was no significant difference inthe rates of BPD [typical RR 0.91 (0.66, 1.24), RD -0.02 (-0.08,0.04)]. The trend towards lower rates of BPD using a longer ITin the studies of Spahr 1980, Heicher 1981 and OCTAVE 1991is offset by their higher rate of death.For all primary outcomes there was no significant heterogeneityof treatment effect (I2 = 0 for all outcomes)Secondary OutcomesOxygenationImprovements in oxygenation using our pre-defined criteria(AaDO2, OI, P/F ratios) were not reported when comparing ITgreater than O.5 seconds with shorter IT. Spahr 1980 providesdata for the mean AaDO2 (SD) after 6 hours following the inter-vention (Table 4). At this early time point, no statistically signifi-cant differences were noted. Analysing survivors only, the AaDO2was significantly lower at 24 and 48 hours in infants ventilatedusing an IT of 2 seconds versus 1 second.
Duration of mechanical ventilation and oxygen therapyData on duration of ventilation in the included studies were ei-ther incomplete or presented in a format not permitting a pooledanalysis. Pohlandt 1992 and Greenough 1989 did not publish suf-ficient data to allow a comparison between groups with respectto duration of mechanical ventilation and oxygen therapy. Spahr1980 presented data only from survivors in the first week of lifeand Heicher 1981 excluded more than 10% of babies from theanalysis of these outcomes. OCTAVE 1991 reported no signifi-cant difference between groups with respect to median age of ex-tubation and median age at weaning from supplemental oxygen.
Neurological injuryNo individual study reported the outcomes of IVH or PVL basedon ultrasound findings. Spahr 1980 reported the incidence of IVHthrough examination at autopsy in those who died and using com-puted tomography in two survivors. Heicher 1981 described largeIVH diagnosed at autopsy but rates of IVH in survivors was notreported. It was the intention of the OCTAVE 1991 group to re-port on rates of IVH but difficulty was found in retrieving imagesfrom all the participating centres and this was replaced by longterm neurodevelopmental follow up. In the available data fromtwo trials, no significant effect on this outcome was shown (Table6).Patent ductus arteriosus
Symptomatic patent ductus arteriosus (PDA) requiring treatmentwas reported in the studies of Spahr and Pohlandt. There was aconsiderable difference in the rates of PDA, with Spahr (conductedin 1978) reporting an incidence of 9% in all recruited infants
7Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A77
and Pohlandt (conducted between 1983 and 1985) reporting anincidence of 35% of all recruited infants. This probably reflectsthe improvements in echocardiography in the intervening period.In a pooled analysis of these two trials, no significant effect on thisoutcome was shown (Table 6)Developmental delayNo individual study used formal psychometric testing to assesslong term developmental status. The single study reporting longterm outcome data (OCTAVE 1991) followed up surviving infantsunder 33 weeks gestation to a median age of 18 months. Thetools used for these assessments were not stated. The outcomesreported were cerebral palsy (Table 7), sensory deficits (impairedhearing or vision; Tables 8 and 9) or global developmental delay.There was an increased rate of cerebral palsy with small numbers(12 versus 4 cases) in infants treated with long IT but this was ofborderline statistical significance [RR 2.90 (0.97, 8.65), RD 0.09(0.00, 0.17).
Subgroup Analyses
Comparison 02: Long (>0.5 seconds) vs Short (<0.5 seconds)IT
This subgroup analysis involved all the included studies except forSpahr 1980 and resulted in findings similar to the overall analysis.In a pooled analysis of four trials involving 625 infants, mortalitybefore hospital discharge showed a trend disfavouring a long ITwhich did not reach statistical significance [typical RR 1.24 (0.96,-1.60), typical RD 0.06 (-0.01, 0.13)]. A long IT was associatedwith a significant increase in the number of air leaks [typical RR1.56 (1.24, 1.97), typical RD 0.13 (0.07,0.20), NNT 8 (5,14)].There was no significant difference in rates of BPD [typical RR0.92 (0.66, 1.28)].Analysis based on the overall ventilator strategya. constant IT - only Heicher 1981 maintained IT at 1.0 vs 0.5seconds in the acute and weaning period.b. constant I/E ratios - only Spahr 1980 and OCTAVE 1991 aimedto keep I:E ratios constant. It should be noted that the ITs in thesetwo studies are vastly different with the duration of inspiration setin the “short IT” arm of Spahr 1980 equal to that set in the “longIT” arm of OCTAVE 1991. With the vastly different ITs andventilator rates and thus very different absolute ratios, a subgroupanalysis of these two studies was not performed based on ventilatorstrategy. These two trials were combined using all enrolled infantsin Spahr 1980 and those with HMD in OCTAVE 1991 in a pooledanalysis of infants with HMD (see below).c. where both IT and I/E ratios are variedIn all studies the short IT remained unchanged. In most of thestudies adjustments of the long IT were permissible when weaningcommenced. Variable I:E ratios were seen in the long IT groupwhen rate was either increased or decreased. The ventilator rateranged between 20 and 40 for Heicher 1981 and OCTAVE 1991
and 30 to 40 for Pohlandt 1992. Thus a subgroup analysis wasnot performed.Comparison 03: Long versus short IT in HMD (subgroup anal-ysis by diagnosis)Although the primary respiratory diagnosis was HMD in most in-fants in the included studies, only Spahr 1980 and OCTAVE 1991provided published data on outcomes by respiratory pathology. Ina pooled analysis of patients with HMD, a significant increase inmortality before discharge [typical RR 1.54 (1.06, 2.23), typicalRD 0.12 (0.02, 0.21), NNT 8 (5, 50)] and air leak [typical RR1.73 (1.17, 2.57), typical RD 0.14 (0.04, 0.24), NNT 7 (4, 25)]was seen in the infants randomised to a long IT.Gestational ageAll the trials except for Pohlandt 1992 (infants less than or equal to32 weeks gestation randomised) had no restrictions on gestationalage at recruitment. In this individual study there was a statisticallysignificant increase in air leak in the long IT group. No significantdifferences were seen in rates of mortality before hospital dischargeand BPD.Comparison 04: Long versus short IT (subgroup analysis oftrials allowing use of muscle relaxants)Muscle relaxants were allowed in the trials of Heicher 1981,Pohlandt 1992 and OCTAVE 1991 but not used in Spahr 1980and Greenough 1989. Data on the extent of use of muscle relax-ants were not presented in the report by Pohlandt 1992, and wereincomplete in OCTAVE 1991 (published data was sourced fromonly of one of the participating centres). Similar use of musclerelaxants occurred in each arm of Heicher 1981 (13 versus 14). Inan analysis confined to the three trials permitting muscle relaxantsusage, there was a significant increase in air leak [typical RR 1.56(1.25, 1.94), typical RD 0.14 (0.07, 0.21), NNT 7 (5, 14)] andan increase in mortality [typical RR 1.26 (0.97, 1.62), typical RD0.07 (-0.01, 0.14)] which reached marginal statistical significance,in infants managed with a long IT. In the trials where no musclerelaxants were used (Spahr 1980, Greenough 1989), there were nostatistically significant differences in air leak or mortality.Trials using surfactantNone of the trials used surfactant in the management of HMD.
Ventilator modeAll infants were ventilated using continuous mandatory ventila-tion. No synchronised modes were used. Ventilator types differed(see table of included studies).
D I S C U S S I O N
The goals of mechanical ventilation are to achieve and maintainoxygenation and remove carbon dioxide whilst minimizing therisk of lung injury. These studies were performed at a time whenmortality rates for ventilated infants with hypoxic respiratory fail-ure were considerably higher than current rates (Doyle 1999). Ma-
8Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A78
jor changes since their publication may reduce the applicability ofthe results of these meta- analyses. These changes include the rou-tine use of antenatal steroids and postnatal surfactant. The majorcomplications of HMD seen at the time of these studies (namelyhypoxia and air leak despite intervening with IPPV) have been sig-nificantly reduced by surfactant replacement therapy (Soll 2004).Although both these interventions have significantly altered thetreatment of HMD, the risk of mortality has been markedly re-duced but has not been removed. Morbidities such as air leakand BPD continue to occur. The classical definition of BPD byNorthway 1967 was based on larger and more mature (>30 weeks)newborns and has been replaced by oxygen and/or ventilator de-pendency at 36 weeks post conceptual age to reflect the presentpopulation of newborn infants with HMD. These extremely lowbirth weight, extremely premature infants continue to have signif-icantly high rates of BPD. None of the studies in this systematicreview looked at ventilatory or oxygen requirements at 36 weeks.
The use of long vs short ITs has not been well studied in subjectswith established BPD or evolving BPD. The one study excludedfrom this meta- analysis (Nilmeier 1995) investigated the effectsof a long or short IT on subjects with BPD (more than one monthof age and ventilated) and looked at pulmonary mechanics andoxygenation before and after one week of the intervention. Al-though a long IT improved delivered tidal volumes, no significantdifferences were seen in duration of ventilation or hospital stay.
The IT set by the attending clinician is a single parameter thatcan influence oxygenation. Other manipulations include provid-ing sufficient PEEP, so that positive pressure delivered through-out the respiratory cycle is above the critical opening pressure ofthe lung, thus avoiding repeated inflation deflation cycles whichare injurious to the lung. Gas flow rates are usually held constantand the effects of varying gas flow rates on oxygenation and BPDhave yet to be investigated. Of all the included studies, only one(Spahr 1980) provided data on the effects of IT on oxygenationin the acute phase of HMD. Whilst the study used ITs in botharms that would be considered “long”, this was a study primar-ily comparing two I:E ratios. The improvements in oxygenationseen in inverse ratio ventilation can be attained (using the samephysiological principles of alveolar recruitment and maintenanceof lung volumes) if appropriate PEEP is applied throughout therespiratory cycle. The optimum level of PEEP will vary accordingto the underlying pulmonary pathophysiology and is yet to beevaluated in a formal clinical trial.
The range of ITs used in the long IT groups in this systematicreview was 0.66 to 2.0 seconds. In acute HMD, where the medianIT is 0.3 seconds (Ahluwalia 1994), these ITs are likely to causeasynchrony and patient discomfort. If these long ITs were used insurfactant treated infants, the combination of improved compli-ance and greater likelihood of active expiration against a positivepressure inflation may increase the risk of air leak. Thus using along IT in acute HMD following surfactant replacement may be
an unnecessary strategy unless faced with continuing severe hy-poxia. In these situations there are current alternative modes oftherapy to improve oxygenation if conventional ventilation is fail-ing, including rescue high frequency ventilation and the use of in-haled nitric oxide in mature infants. Perinatal units are now widelyusing ventilators that synchronise with the infant’s breathing andtarget tidal volumes. With synchronised ventilation, square waveventilation has been replaced by sinusoidal waves so that meanairway pressures are less influenced by the length of the IT.
This review has shown no advantage in using long IT over shortIT in the treatment of acute respiratory failure (mainly HMD).Whilst there is increasing use of non invasive ventilation such asnasal continuous positive airway pressure to avoid ventilator in-duced lung injury in the acute management of HMD, mechanicalventilation will continue to have a role in extremely immature in-fants and those with HMD complicated by apnea. In institutionswhere surfactant is unavailable, lengthening the IT may improveoxygenation in the acute phase. However, as compliance improves,the IT should be reviewed regularly as this review demonstratesa significant risk of lung injury (airway leak) with a long IT. Theuse of a long IT where time constants are longer than acute HMDsuch as premature infants with BPD, meconium aspiration syn-drome and newborns in cardiac failure may be appropriate and isyet to be investigated.
Whilst neonatal respiratory failure is not a single disease entity,relatively recent advances in technology have allowed the bedsidedisplay of real time inspiratory and expiratory tidal volumes. Clin-icians therefore have the ability to adjust the IT as pulmonarymechanics change during the course of an infant’s illness. Thesestrategies will need to be evaluated in future clinical trials.
A U T H O R S ’ C O N C L U S I O N S
Implications for practice
Long inspiratory times when used in acute HMD in a populationnot exposed to antenatal steroids and postnatal surfactant are as-sociated with higher rates of mortality and morbidity. Stiff lungswith HMD have very short time constants. Mechanically venti-lated infants with HMD and especially those treated in institu-tions where these adjunctive therapies are not available should beventilated using a short IT.
Implications for research
The availability of real time, continuous measurements of pul-monary mechanics may enhance clinicians’ ability to detect theoptimal IT for infants with different underlying pathologies andat different time points in their illness. Future research should ex-amine whether new monitoring equipment and ventilator strate-gies, including adjusting ITs to match underlying compliance, can
9Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A79
further reduce the harmful effects of neonatal mechanical ventila-tion.
R E F E R E N C E S
References to studies included in this review
Greenough 1989 published data only
Greenough A, Pool J, Gamsu H. Randomised controlled trial oftwo methods of weaning from high frequency positive pressure
ventilation. Archives of Disease in Childhood 1989;64:834–8.
Heicher 1981 published data onlyHeicher DA, Kasting DS, Harrod JR. Prospective clinical
comparison of two methods for mechanical ventilation of neonates:Rapid rate and short inspiratory time versus slow rate and long
inspiratory time. Journal of Pediatrics 1981;98:957–61.
OCTAVE 1991 published data only
Multicentre randomised controlled trial of high frequency positivepressure ventilation. Oxford Region Controlled Trial of Artificial
Ventilation OCTAVE Study Group. Archives of Disease inChildhood 1991;66:770–5.
Pohlandt 1992 published data only
Pohlandt F, Saule H, Schroder H, Leonhardt A, Hornchen H,Wolff C, et al.Decreased incidence of extra-alveolar air leakage or
death prior to air leakage in high versus low rate positive pressureventilation: results of a randomised seven-centre trial in preterm
infants. European Journal of Pediatrics 1992;151:904–9.
Spahr 1980 published data only
Spahr RC, Klein AM, Brown DR, MacDonald HM, Holzman IR.Hyaline membrane disease - A controlled study of inspiratory to
expiratory ratio in its manaagement by ventilator. American Journalof Diseases of Children 1980;134:373–6.
References to studies excluded from this review
Nilmeier 1995 published data onlyNilmeier NL, Hodge GB, Dunn CE, Benitz WE, Ariagno RL. One
second inspiratory time improves lung mechanics in preterminfants who are ventilator dependent post respiratory distress
syndrome. Pediatric Research 1995;37:344A.
Additional references
Adamson 1968
Adamson TM, Collins LM, Dehan M, Hawker JM, Reynolds EO,Strang LB. Mechanical ventilation in newborn infants with
respiratory failure. Lancet 1968;2:227–31.
Ahluwalia 1994
Ahluwalia JS, Morley CJ, Mockridge JN. Computeriseddetermination of spontaneous inspiratory and expiratory times in
premature neonates during intermittent positive pressureventilation. II: Results from 20 babies. Archives of Disease in
Childhood Fetal Neonatal Edition 1994;71:F161–4.
Boros 1979Boros SJ. Variations in inspiratory:expiratory ratio and airway
pressure wave form during mechanical ventilation: the significanceof mean airway pressure. Journal of Pediatrics 1979;94:114–7.
Boros 1984Boros SJ, Bing DR, Mammel MC, Hagen E, Gordon MJ. Using
conventional infant ventilators at unconventional rates. Pediatrics1984;74:487–92.
Cumarasamy 1973Cumarasamy N, Nussli R, Vischer D, Dangel PH, Duc GV.
Artificial ventilation in hyaline membrane disease: the use ofpositive end-expiratory pressure and continuous postive airway
pressure. Pediatrics 1973;51:629–40.
D-Papadopoulos 1965
Delivoria-Papadopoulos M, Levison H, Swyer PR. Intermittentpositive pressure ventilation as a treatment in severe respiratory
distress syndrome. Arch Dis Child 1965;40:474–9.
Doyle 1999Doyle LW, Gultom E, Chuang SL, James M, Davis P, Bowman E.
Changing mortality and causes of death in infants 23-27 weeks’gestational age. Journal of Paediatrics and Child Health 1999;35:
255–9.
Goldsmith 1996Goldsmith JP, Karotkin E (eds). Assisted Ventilation of the Newborn.
3rd Edition. Philadelphia: Saunders Company, 1996.
Greenough 1986
Greenough A, Morley CJ, Pool J. Fighting the ventilator - are fastrates an effective alternative to paralysis. Early Human Development
1986;13:189–94.
Greenough 1987
Greenough A, Pool J, Greenall F, Morley C, Gamsu H.Comparison of different rates of artificial ventilation in preterm
neonates with respiratory distress syndrome. Acta PaediatricaScandinavica 1987;76:706–12.
Greenough 2004
Greenough A, Milner AD, Dimitriou G. Synchronized mechanicalventilation for respiratory support in newborn infants. Cochrane
Database of Systematic Reviews 2004, Issue 2. [DOI: 10.1002/14651858.CD000456.pub2]
Gregory 1971
Gregory GA, Kitterman JA, Phibbs RH, Tooley WH, HamiltonWK. Treatment of the idiopathic respiratory distress syndrome with
continuous positive airway pressure. New England Journal ofMedicine 1971;284:1333–40.
Harris 1996
Harris TR. In: Goldsmith JP, Karotkin EH editor(s). Assistedventilation of the newborn. 3rd Edition. Philadelphia: W.B
Saunders, 1996.
10Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A80
Herman 1973
Herman S, Reynolds EO. Methods of improving oxygenation ininfants mechanically ventilated for severe hyaline membrane
disease. Archives of Disease in Childhood 1973;48:612–17.
Murdock 1970
Murdock AI, Linsao L, Reid MM, Sutton MD, Tilak KS, UlanOA, Swyer PR. Mechanical ventilation in the respiratory distress
syndrome: a controlled trial. Archives of Disease in Childhood 1970;45:624–33.
Northway 1967Northway WH Jr, Rosan RC, Porter DY. Pulmonary disease
following respirator therapy of hyaline membrane disease.Bronchopulmonary dysplasia. New England Journal of Medicine
1967;276:357–68.
Papile 1978
Papile LA, Burstein J, Burstein R, Koffler H. Incidence andevolution of subependymal and intraventricular hemorrhage: a
study of infants with birthweights less than 1,500 gm. Journal ofPediatrics 1978;92:529–34.
Rennie 1987Rennie J, South M, Morley C. Cerebral blood flow velocity
variability in infants receiving assisted ventilation. Archives ofDisease in Childhood 1987;62:1247–51.
Reynolds 1971Reynolds EOR. Effect of alterations in mechanical ventilator
settings on pulmonary gas exchange in hyaline membrane disease.Archives of Disease in Childhood 1971;46:152–9.
Soll 1992Soll RF, McQueen MC. Respiratory distress syndrome. In: Sinclair
JC, Bracken MB editor(s). Effective Care of the Newborn Infant.
Oxford: Oxford University Press, 1992:325–58.
Soll 2004
Soll RF, Blanco F. Natural surfactant versus synthetic surfactant forneonatal respiratory distress syndrome. Cochrane Database of
Systematic Reviews 2004, Issue 2. [DOI: 10.1002/14651858.CD000144]
Stewart 1981Stewart AR, Finer NN, Peters KL. Effects of alterations of
inspiratory and expiratory pressures and inspiratory/expiratoryratios on mean airway pressure, blood gases, and intracranial
pressure. Pediatrics 1981;67:474–81.
Upton 1990
Upton CJ, Milner AD, Stokes GM. The effect of changes ininspiratory time on neonatal triggered ventilation. European Journal
of Pediatrics 1990;149:648–50.
Usher 1963Usher R. Reduction of mortality from respiratory distress syndrome
of prematurity with early administration of intravenous glucose andsodium bicarbonate. Pediatrics 1963;32:966–75.
Weigl 1973Weigl, J. The infant lung. A case against high respiratory rates in
controlled neonatal ventilation. Respiratory Therapy 1973;3:57.
Yost 2003
Yost CC, Soll RF. Early versus delayed selective surfactant treatmentfor neonatal respiratory distress syndrome. Cochrane Database of
Systematic Reviews 2003, Issue 3. [DOI: 10.1002/14651858.CD001456]
∗ Indicates the major publication for the study
11Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A81
C H A R A C T E R I S T I C S O F S T U D I E S
Characteristics of included studies [ordered by study ID]
Greenough 1989
Methods Concealment of allocation: Can’t tell - method of randomisation to either long IT (group A) or short IT(group B) not stated.Blinding of intervention: NoCompleteness of follow up: Yes for short-term outcomes - all participants accounted for. Blinding ofoutcome measurement: NoSingle centre study. Method of randomisation not stated. Intervention not blinded. Follow up completeto stated primary outcome (duration of weaning).
Participants Single centre study. All infants with RDS (n=40) and ventilated with HFPPV (>60bpm) were recruitedwhen weaning commenced (PIP = 20cm H20 and rate =60bpm).
Interventions At recruitment all infants were ventilated using an IT of 0.5 seconds. Weaning then commenced usingone of two strategies. In Group A (long IT) (n=20), the rate was reduced by increasing both the IT andET maintaining a fixed I:E ratio at 1:1.2. Maximum IT in this group was 1 second. In group B (shortIT) (n=20), the rate was weaned by keeping the IT fixed at 0.5 secs whilst increasing the expiratory time(ET). Single (Sechrist) type of ventilator.
Outcomes Air leakIndices of oxygenation (paO2)Duration of weaningRate of reintubation
Notes Year of study = 1989. Pre-surfactant. Age at recruitment not specified.
Risk of bias
Item Authors’ judgement Description
Allocation concealment? Unclear B - Unclear
Heicher 1981
Methods Concealment of allocation: No - quasi-randomised using alternate allocation for long and short ITsBlinding of intervention: NoCompleteness of follow up: Yes for short-term outcomes (all outcomes complete for hospital stay)Blinding of outcome measurement: No
Participants All infants >750g and ventilated with abnormal CXR findings (n=102). Infants with gross abnormalities,chromosomal anomalies and meconium aspiration were excluded.
Interventions Long IT (n=51) compared with short IT (n=51). Long IT= 1.0 seconds; rates used 20-40bpm. ShortIT=0.5 seconds; rate used 60bpm. Single ventilator (Babybird) used in both groups. Criteria for failedtreatment and relaxation of allocated ventilator strategy were as follows; 1. Hypoxia - either (i)PaO2 < 50
12Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A82
Heicher 1981 (Continued)
torr with FiO2 of 1.0 and PEEP of 8 to 10cm H20 or (ii) Fi02 > 0.6 (short IT) or PIP > 30cm H20 (longIT) for 72 hours. 2. Hypercarbia - PaCO2 > 45 torr on maximum PIP and rate permitted for the group.
Outcomes MortalityAir leakBPD at 30 days by Northway criteria.Combined outcome of death and IVH
Notes Years of study = 1978 - 79. Pre-surfactant. Age at recruitment not specified.
Risk of bias
Item Authors’ judgement Description
Allocation concealment? No C - Inadequate
OCTAVE 1991
Methods Concealment of allocation: Yes - sealed opaque, serially numbered envelopes stored at co-ordinatingcentre.Blinding of intervention: NoCompleteness of follow up: Yes for short-term outcomes -100% follow up for in-hospital outcomes with 97% of infants born less than 33 weeks accounted forin the long term outcome assessment at 18 months. Blinding of outcome measurement: not for primaryoutcomes. 98% of neurodevelopmental follow up completed by physicians blinded to ventilation strategy
Participants Newborns of any gestation ventilated for any reason (n=346). Infants with meconium aspiration excluded.Infants were enrolled before 72 hours of age.
Interventions Long IT (n=172) compared with short IT (n=174). Long IT= 1.0 seconds; starting rate=20bpm (range=20-40bpm)and I:E ratios not fixed. Short IT=0.33 seconds; rate used =60bpm and I:E ratio fixed at 1:2.Single ventilator (Sechrist IV 100B) used in both groups. Dopamine and tolazoline used at physician’sdiscretion to treat hypotension or reduce right to left shunting. The use of muscle relaxants was permittedfor recurrent air leaks and for poor gas exchange despite recruitment manouvres and optimisation ofhaemodynamics. Criteria for failed treatment and relaxation of allocated ventilator strategy were as follows;1. Hypoxia - PaO2 < 50mmHg despite FiO2 >0.95 and MAP of 17cmH20. 2. Recurrent pneumothoraces
Outcomes Death before dischargeAir leakBPD defined as need for supplemental oxygen at 28 daysA subgroup of infants less than 33 weeks gestation had long term follow up at a median age of 18 months
Notes Years of study = 1983 - 86. Pre-surfactant.
Risk of bias
Item Authors’ judgement Description
13Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A83
OCTAVE 1991 (Continued)
Allocation concealment? Yes A - Adequate
Pohlandt 1992
Methods Concealment of allocation: No - all ventilated infants provisionally randomised using one of three ran-domisation tables (stratified by gestational age). If infant subsequently excluded they were replaced by thenext preterm infant in the same gestational age category.Blinding of intervention: NoCompleteness of follow up: Yes for primary outcomes - 100% follow up for in-hospital outcomesBlinding of outcome measurement: Only for diagnosis of air leak and BPD
Participants Preterm infants equal to or less than 32 weeks gestation and mechanically ventilated (n=181)were ran-domised if FiO2 >0.4 after 3 hours of birth (no upper age limit given). Infants later excluded if the FiO2was < 0.4 to maintain PaO2 >50mmHg or if there were any violations of allocated ventilation protocol (exclusions post randomisation=44)
Interventions Long IT (n=63) compared with short IT (n=74). Long IT=1.0 seconds; rates of 30 to 40bpm; I:E ratio1:1 increased to 2:1 in case of hypoxaemia. IT reduced to 0.66 seconds during weaning. Short IT=0.33seconds; rate=60bpm. I:E ratios kept constant at 1:2; weaned by reducing PIP, PEEP and inspired oxygen.Seven different ventilators were used in this study. No failure criteria described in the report.
Outcomes MortalityExtra-alveolar air leakBPD at 28 days by Northway criteria
Notes Years of study = 1983 - 85. Parental consent not sought as ventilator frequencies of 30 -100 widely usedin the participating units. Pre-surfactant.
Risk of bias
Item Authors’ judgement Description
Allocation concealment? No C - Inadequate
Spahr 1980
Methods Concealment of allocation: No - randomised by coin tossBlinding of intervention: NoCompleteness of follow up: Yes for short-term outcomesBlinding of outcome measurement: No
Participants All infants admitted in a 12 month period and ventilated for HMD (n=74). Five post randomisationexclusions (pre-existing air leaks=3, sepsis=1, use of paralysing agent=1)
Interventions Long IT (n=36) compared with short IT (n=33). Long IT=2 seconds; rate=20bpm; I:E ratio 2:1. ShortIT=1 seconds; rate=20bpm; I:E ratio 1:2. Target PaO2 45 - 60mmHg and pH 7.25 to 7.35. Single
14Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A84
Spahr 1980 (Continued)
ventilator (Babybird). No failure criteria described in the report.
Outcomes Mortality at 28 daysAir leakBPD defined as need for supplemental oxygen at one month with typical radiographic findingsPulmonary haemorrhage, IVH (all grades), PDA.
Notes Year of study = 1978. Pre-surfactant and pre antenatal steroids.
Risk of bias
Item Authors’ judgement Description
Allocation concealment? No C - Inadequate
HMD = Hyaline Membrane Disease, RDS = Respiratory Distress Syndrome, IT = inspiratory time, IVH = intraventricular haemorrhage,HFPPV = High Frequency Positive Pressure Ventilation
Characteristics of excluded studies [ordered by study ID]
Nilmeier 1995 Randomised controlled trial of infants recruited at two weeks of age. Oxygenation and respiratory mechanics werecompared after a week of intervention at two ITs (1.0 and 0.4 seconds). Published results however were incomplete(data from only one group) and descriptive in nature ( not using our prespecified indices of AaDO2 and PF ratios).
15Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A85
D A T A A N D A N A L Y S E S
Comparison 1. Long vs Short IT as defined by investigators (all trials)
Outcome or subgroup titleNo. ofstudies
No. ofparticipants Statistical method Effect size
1 Mortality before discharge 5 694 Risk Ratio (M-H, Fixed, 95% CI) 1.26 [1.00, 1.59]2 Air leak 5 685 Risk Ratio (M-H, Fixed, 95% CI) 1.56 [1.25, 1.94]
3 BPD (supplemental oxygen at28 days)
4 654 Risk Ratio (M-H, Fixed, 95% CI) 0.91 [0.66, 1.24]
4 AaDO2 (mmHg) values after 6hours of intervention
1 69 Mean Difference (IV, Fixed, 95% CI) -18.60 [-93.78,56.58]
5 IVH (all grades) 2 171 Risk Ratio (M-H, Fixed, 95% CI) 1.11 [0.72, 1.71]6 Patent ductus arteriosus (PDA) 2 206 Risk Ratio (M-H, Fixed, 95% CI) 0.91 [0.58, 1.43]
7 Cerebral palsy in survivors lessthan 33 weeks gestation at birth
1 177 Risk Ratio (M-H, Fixed, 95% CI) 2.9 [0.97, 8.65]
8 Visual impairment in survivorsless than 33 weeks gestation atbirth
1 177 Risk Ratio (M-H, Fixed, 95% CI) 2.09 [0.83, 5.26]
9 Sensorineural hearing loss insurvivors less than 33 weeksgestation at birth
1 177 Risk Ratio (M-H, Fixed, 95% CI) 1.93 [0.60, 6.19]
Comparison 2. Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds)
Outcome or subgroup titleNo. ofstudies
No. ofparticipants Statistical method Effect size
1 Mortality before discharge 4 625 Risk Ratio (M-H, Fixed, 95% CI) 1.24 [0.96, 1.60]2 Air leak 4 616 Risk Ratio (M-H, Fixed, 95% CI) 1.56 [1.24, 1.97]
3 BPD (supplemental oxygenationat 28 days)
3 585 Risk Ratio (M-H, Fixed, 95% CI) 0.92 [0.66, 1.28]
4 IVH (all grades) 1 102 Risk Ratio (M-H, Fixed, 95% CI) 1.08 [0.55, 2.14]
16Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A86
Comparison 3. Long vs Short IT in HMD (subgroup analysis by diagnosis)
Outcome or subgroup titleNo. ofstudies
No. ofparticipants Statistical method Effect size
1 Mortality before hospitaldischarge
2 306 Risk Ratio (M-H, Fixed, 95% CI) 1.54 [1.06, 2.23]
2 Air leak 2 303 Risk Ratio (M-H, Fixed, 95% CI) 1.73 [1.17, 2.57]
3 BPD (supplemental oxygen at28 days)
2 253 Risk Ratio (M-H, Fixed, 95% CI) 0.88 [0.60, 1.30]
Comparison 4. Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation)
Outcome or subgroup titleNo. ofstudies
No. ofparticipants Statistical method Effect size
1 Mortality before discharge 5 694 Risk Ratio (M-H, Fixed, 95% CI) 1.26 [1.00, 1.59]1.1 Muscle relax allowed 3 585 Risk Ratio (M-H, Fixed, 95% CI) 1.26 [0.97, 1.62]1.2 No muscle relaxation 2 109 Risk Ratio (M-H, Fixed, 95% CI) 1.26 [0.73, 2.18]
2 Air leak 5 685 Risk Ratio (M-H, Fixed, 95% CI) 1.56 [1.25, 1.94]2.1 Muscle relax allowed 3 576 Risk Ratio (M-H, Fixed, 95% CI) 1.54 [1.22, 1.94]2.2 No muscle relaxation 2 109 Risk Ratio (M-H, Fixed, 95% CI) 1.69 [0.82, 3.47]
3 BPD (supplemental oxygen at28 days)
4 654 Risk Ratio (M-H, Fixed, 95% CI) 0.91 [0.66, 1.24]
3.1 Muscle relax allowed 3 585 Risk Ratio (M-H, Fixed, 95% CI) 0.92 [0.66, 1.28]3.2 No muscle relaxation 1 69 Risk Ratio (M-H, Fixed, 95% CI) 0.73 [0.21, 2.50]
4 IVH (all grades) 2 171 Risk Ratio (M-H, Fixed, 95% CI) 1.11 [0.72, 1.71]4.1 Muscle relax allowed 1 102 Risk Ratio (M-H, Fixed, 95% CI) 1.08 [0.55, 2.14]4.2 No muscle relaxation 1 69 Risk Ratio (M-H, Fixed, 95% CI) 1.13 [0.65, 1.97]
17Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A87
Analysis 1.1. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 1 Mortality
before discharge.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 1 Mortality before discharge
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Greenough 1989 0/20 1/20 1.7 % 0.33 [ 0.01, 7.72 ]
Heicher 1981 17/51 14/51 15.7 % 1.21 [ 0.67, 2.19 ]
OCTAVE 1991 50/172 41/174 45.8 % 1.23 [ 0.86, 1.76 ]
Pohlandt 1992 25/63 22/74 22.7 % 1.33 [ 0.84, 2.12 ]
Spahr 1980 18/36 12/33 14.1 % 1.38 [ 0.79, 2.40 ]
Total (95% CI) 342 352 100.0 % 1.26 [ 1.00, 1.59 ]Total events: 110 (Long IT), 90 (Short IT)
Heterogeneity: Chi2 = 0.87, df = 4 (P = 0.93); I2 =0.0%
Test for overall effect: Z = 1.95 (P = 0.052)
0.02 0.1 1 10 50
Favours Long IT Favours Short IT
Analysis 1.2. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 2 Air leak.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 2 Air leak
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Greenough 1989 2/20 0/20 0.6 % 5.00 [ 0.26, 98.00 ]
Heicher 1981 18/51 7/51 8.6 % 2.57 [ 1.18, 5.62 ]
OCTAVE 1991 44/167 32/170 38.9 % 1.40 [ 0.94, 2.09 ]
Pohlandt 1992 46/63 37/74 41.7 % 1.46 [ 1.11, 1.92 ]
Spahr 1980 13/36 8/33 10.2 % 1.49 [ 0.71, 3.13 ]
Total (95% CI) 337 348 100.0 % 1.56 [ 1.25, 1.94 ]Total events: 123 (Long IT), 84 (Short IT)
Heterogeneity: Chi2 = 2.67, df = 4 (P = 0.62); I2 =0.0%
Test for overall effect: Z = 3.95 (P = 0.000077)
0.02 0.1 1 10 50
Favours Long IT Favours Short IT
18Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A88
Analysis 1.3. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 3 BPD
(supplemental oxygen at 28 days).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 3 BPD (supplemental oxygen at 28 days)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Heicher 1981 4/51 5/51 7.8 % 0.80 [ 0.23, 2.81 ]
OCTAVE 1991 38/172 42/174 65.3 % 0.92 [ 0.62, 1.34 ]
Pohlandt 1992 11/63 13/74 18.7 % 0.99 [ 0.48, 2.06 ]
Spahr 1980 4/36 5/33 8.2 % 0.73 [ 0.21, 2.50 ]
Total (95% CI) 322 332 100.0 % 0.91 [ 0.66, 1.24 ]Total events: 57 (Long IT), 65 (Short IT)
Heterogeneity: Chi2 = 0.22, df = 3 (P = 0.97); I2 =0.0%
Test for overall effect: Z = 0.61 (P = 0.54)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
Analysis 1.4. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 4 AaDO2(mmHg) values after 6 hours of intervention.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 4 AaDO2 (mmHg) values after 6 hours of intervention
Study or subgroup Long IT Short IT Mean Difference Weight Mean Difference
N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI
Spahr 1980 36 282.3 (156) 33 300.9 (162) 100.0 % -18.60 [ -93.78, 56.58 ]
Total (95% CI) 36 33 100.0 % -18.60 [ -93.78, 56.58 ]Heterogeneity: not applicable
Test for overall effect: Z = 0.48 (P = 0.63)
-50 -25 0 25 50
Favours Long IT Favours Short IT
19Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A89
Analysis 1.5. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 5 IVH (all
grades).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 5 IVH (all grades)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Heicher 1981 13/51 12/51 46.9 % 1.08 [ 0.55, 2.14 ]
Spahr 1980 16/36 13/33 53.1 % 1.13 [ 0.65, 1.97 ]
Total (95% CI) 87 84 100.0 % 1.11 [ 0.72, 1.71 ]Total events: 29 (Long IT), 25 (Short IT)
Heterogeneity: Chi2 = 0.01, df = 1 (P = 0.93); I2 =0.0%
Test for overall effect: Z = 0.46 (P = 0.65)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
Analysis 1.6. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 6 Patent
ductus arteriosus (PDA).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 6 Patent ductus arteriosus (PDA)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Pohlandt 1992 20/63 28/74 92.5 % 0.84 [ 0.53, 1.34 ]
Spahr 1980 4/36 2/33 7.5 % 1.83 [ 0.36, 9.36 ]
Total (95% CI) 99 107 100.0 % 0.91 [ 0.58, 1.43 ]Total events: 24 (Long IT), 30 (Short IT)
Heterogeneity: Chi2 = 0.83, df = 1 (P = 0.36); I2 =0.0%
Test for overall effect: Z = 0.40 (P = 0.69)
0.1 0.2 0.5 1 2 5 10
Favours Long IT Favours Short IT
20Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A90
Analysis 1.7. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 7 Cerebral
palsy in survivors less than 33 weeks gestation at birth.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 7 Cerebral palsy in survivors less than 33 weeks gestation at birth
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
OCTAVE 1991 12/90 4/87 100.0 % 2.90 [ 0.97, 8.65 ]
Total (95% CI) 90 87 100.0 % 2.90 [ 0.97, 8.65 ]Total events: 12 (Long IT), 4 (Short IT)
Heterogeneity: not applicable
Test for overall effect: Z = 1.91 (P = 0.056)
0.2 0.5 1 2 5
Favours Long IT Favours Short IT
Analysis 1.8. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 8 Visualimpairment in survivors less than 33 weeks gestation at birth.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 8 Visual impairment in survivors less than 33 weeks gestation at birth
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
OCTAVE 1991 13/90 6/87 100.0 % 2.09 [ 0.83, 5.26 ]
Total (95% CI) 90 87 100.0 % 2.09 [ 0.83, 5.26 ]Total events: 13 (Long IT), 6 (Short IT)
Heterogeneity: not applicable
Test for overall effect: Z = 1.57 (P = 0.12)
0.2 0.5 1 2 5
Favours Long IT Favours Short IT
21Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A91
Analysis 1.9. Comparison 1 Long vs Short IT as defined by investigators (all trials), Outcome 9
Sensorineural hearing loss in survivors less than 33 weeks gestation at birth.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 1 Long vs Short IT as defined by investigators (all trials)
Outcome: 9 Sensorineural hearing loss in survivors less than 33 weeks gestation at birth
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
OCTAVE 1991 8/90 4/87 100.0 % 1.93 [ 0.60, 6.19 ]
Total (95% CI) 90 87 100.0 % 1.93 [ 0.60, 6.19 ]Total events: 8 (Long IT), 4 (Short IT)
Heterogeneity: not applicable
Test for overall effect: Z = 1.11 (P = 0.27)
0.2 0.5 1 2 5
Favours Long IT Favours Short IT
Analysis 2.1. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds),Outcome 1 Mortality before discharge.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds)
Outcome: 1 Mortality before discharge
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Greenough 1989 0/20 1/20 2.0 % 0.33 [ 0.01, 7.72 ]
Heicher 1981 17/51 14/51 18.3 % 1.21 [ 0.67, 2.19 ]
OCTAVE 1991 50/172 41/174 53.3 % 1.23 [ 0.86, 1.76 ]
Pohlandt 1992 25/63 22/74 26.5 % 1.33 [ 0.84, 2.12 ]
Total (95% CI) 306 319 100.0 % 1.24 [ 0.96, 1.60 ]Total events: 92 (Long IT), 78 (Short IT)
Heterogeneity: Chi2 = 0.77, df = 3 (P = 0.86); I2 =0.0%
Test for overall effect: Z = 1.66 (P = 0.098)
0.02 0.1 1 10 50
Favours Long IT Favours Short IT
22Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A92
Analysis 2.2. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds),
Outcome 2 Air leak.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds)
Outcome: 2 Air leak
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Greenough 1989 2/20 0/20 0.7 % 5.00 [ 0.26, 98.00 ]
Heicher 1981 18/51 7/51 9.6 % 2.57 [ 1.18, 5.62 ]
OCTAVE 1991 44/167 32/170 43.3 % 1.40 [ 0.94, 2.09 ]
Pohlandt 1992 46/63 37/74 46.5 % 1.46 [ 1.11, 1.92 ]
Total (95% CI) 301 315 100.0 % 1.56 [ 1.24, 1.97 ]Total events: 110 (Long IT), 76 (Short IT)
Heterogeneity: Chi2 = 2.68, df = 3 (P = 0.44); I2 =0.0%
Test for overall effect: Z = 3.82 (P = 0.00013)
0.01 0.1 1 10 100
Favours Long IT Favours Short IT
Analysis 2.3. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds),Outcome 3 BPD (supplemental oxygenation at 28 days).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds)
Outcome: 3 BPD (supplemental oxygenation at 28 days)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Heicher 1981 4/51 5/51 8.5 % 0.80 [ 0.23, 2.81 ]
OCTAVE 1991 38/172 42/174 71.1 % 0.92 [ 0.62, 1.34 ]
Pohlandt 1992 11/63 13/74 20.4 % 0.99 [ 0.48, 2.06 ]
Total (95% CI) 286 299 100.0 % 0.92 [ 0.66, 1.28 ]Total events: 53 (Long IT), 60 (Short IT)
Heterogeneity: Chi2 = 0.09, df = 2 (P = 0.96); I2 =0.0%
Test for overall effect: Z = 0.49 (P = 0.63)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
23Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A93
Analysis 2.4. Comparison 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds),
Outcome 4 IVH (all grades).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 2 Long (greater than 0.5 seconds) vs Short IT (less than or equal to 0.5 seconds)
Outcome: 4 IVH (all grades)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Heicher 1981 13/51 12/51 100.0 % 1.08 [ 0.55, 2.14 ]
Total (95% CI) 51 51 100.0 % 1.08 [ 0.55, 2.14 ]Total events: 13 (Long IT), 12 (Short IT)
Heterogeneity: not applicable
Test for overall effect: Z = 0.23 (P = 0.82)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
Analysis 3.1. Comparison 3 Long vs Short IT in HMD (subgroup analysis by diagnosis), Outcome 1 Mortalitybefore hospital discharge.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 3 Long vs Short IT in HMD (subgroup analysis by diagnosis)
Outcome: 1 Mortality before hospital discharge
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
OCTAVE 1991 32/114 21/123 61.7 % 1.64 [ 1.01, 2.68 ]
Spahr 1980 18/36 12/33 38.3 % 1.38 [ 0.79, 2.40 ]
Total (95% CI) 150 156 100.0 % 1.54 [ 1.06, 2.23 ]Total events: 50 (Long IT), 33 (Short IT)
Heterogeneity: Chi2 = 0.23, df = 1 (P = 0.63); I2 =0.0%
Test for overall effect: Z = 2.29 (P = 0.022)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
24Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A94
Analysis 3.2. Comparison 3 Long vs Short IT in HMD (subgroup analysis by diagnosis), Outcome 2 Air leak.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 3 Long vs Short IT in HMD (subgroup analysis by diagnosis)
Outcome: 2 Air leak
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
OCTAVE 1991 37/112 22/122 71.6 % 1.83 [ 1.16, 2.91 ]
Spahr 1980 13/36 8/33 28.4 % 1.49 [ 0.71, 3.13 ]
Total (95% CI) 148 155 100.0 % 1.73 [ 1.17, 2.57 ]Total events: 50 (Long IT), 30 (Short IT)
Heterogeneity: Chi2 = 0.22, df = 1 (P = 0.64); I2 =0.0%
Test for overall effect: Z = 2.76 (P = 0.0058)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
Analysis 3.3. Comparison 3 Long vs Short IT in HMD (subgroup analysis by diagnosis), Outcome 3 BPD(supplemental oxygen at 28 days).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 3 Long vs Short IT in HMD (subgroup analysis by diagnosis)
Outcome: 3 BPD (supplemental oxygen at 28 days)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
OCTAVE 1991 27/82 37/102 86.3 % 0.91 [ 0.61, 1.36 ]
Spahr 1980 4/36 5/33 13.7 % 0.73 [ 0.21, 2.50 ]
Total (95% CI) 118 135 100.0 % 0.88 [ 0.60, 1.30 ]Total events: 31 (Long IT), 42 (Short IT)
Heterogeneity: Chi2 = 0.11, df = 1 (P = 0.74); I2 =0.0%
Test for overall effect: Z = 0.63 (P = 0.53)
0.2 0.5 1 2 5
Favours Long IT Favours Short IT
25Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A95
Analysis 4.1. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation),
Outcome 1 Mortality before discharge.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation)
Outcome: 1 Mortality before discharge
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
1 Muscle relax allowed
Heicher 1981 17/51 14/51 15.7 % 1.21 [ 0.67, 2.19 ]
OCTAVE 1991 50/172 41/174 45.8 % 1.23 [ 0.86, 1.76 ]
Pohlandt 1992 25/63 22/74 22.7 % 1.33 [ 0.84, 2.12 ]
Subtotal (95% CI) 286 299 84.2 % 1.26 [ 0.97, 1.62 ]Total events: 92 (Long IT), 77 (Short IT)
Heterogeneity: Chi2 = 0.09, df = 2 (P = 0.96); I2 =0.0%
Test for overall effect: Z = 1.76 (P = 0.078)
2 No muscle relaxation
Greenough 1989 0/20 1/20 1.7 % 0.33 [ 0.01, 7.72 ]
Spahr 1980 18/36 12/33 14.1 % 1.38 [ 0.79, 2.40 ]
Subtotal (95% CI) 56 53 15.8 % 1.26 [ 0.73, 2.18 ]Total events: 18 (Long IT), 13 (Short IT)
Heterogeneity: Chi2 = 0.78, df = 1 (P = 0.38); I2 =0.0%
Test for overall effect: Z = 0.84 (P = 0.40)
Total (95% CI) 342 352 100.0 % 1.26 [ 1.00, 1.59 ]Total events: 110 (Long IT), 90 (Short IT)
Heterogeneity: Chi2 = 0.87, df = 4 (P = 0.93); I2 =0.0%
Test for overall effect: Z = 1.95 (P = 0.052)
0.02 0.1 1 10 50
Favours Long IT Favours Short IT
26Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A96
Analysis 4.2. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation),
Outcome 2 Air leak.
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation)
Outcome: 2 Air leak
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
1 Muscle relax allowed
Heicher 1981 18/51 7/51 8.6 % 2.57 [ 1.18, 5.62 ]
OCTAVE 1991 44/167 32/170 38.9 % 1.40 [ 0.94, 2.09 ]
Pohlandt 1992 46/63 37/74 41.7 % 1.46 [ 1.11, 1.92 ]
Subtotal (95% CI) 281 295 89.2 % 1.54 [ 1.22, 1.94 ]Total events: 108 (Long IT), 76 (Short IT)
Heterogeneity: Chi2 = 2.02, df = 2 (P = 0.37); I2 =1%
Test for overall effect: Z = 3.69 (P = 0.00023)
2 No muscle relaxation
Greenough 1989 2/20 0/20 0.6 % 5.00 [ 0.26, 98.00 ]
Spahr 1980 13/36 8/33 10.2 % 1.49 [ 0.71, 3.13 ]
Subtotal (95% CI) 56 53 10.8 % 1.69 [ 0.82, 3.47 ]Total events: 15 (Long IT), 8 (Short IT)
Heterogeneity: Chi2 = 0.62, df = 1 (P = 0.43); I2 =0.0%
Test for overall effect: Z = 1.43 (P = 0.15)
Total (95% CI) 337 348 100.0 % 1.56 [ 1.25, 1.94 ]Total events: 123 (Long IT), 84 (Short IT)
Heterogeneity: Chi2 = 2.67, df = 4 (P = 0.62); I2 =0.0%
Test for overall effect: Z = 3.95 (P = 0.000077)
0.01 0.1 1 10 100
Favours Long IT Favours Short IT
27Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A97
Analysis 4.3. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation),
Outcome 3 BPD (supplemental oxygen at 28 days).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation)
Outcome: 3 BPD (supplemental oxygen at 28 days)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
1 Muscle relax allowed
Heicher 1981 4/51 5/51 7.8 % 0.80 [ 0.23, 2.81 ]
OCTAVE 1991 38/172 42/174 65.3 % 0.92 [ 0.62, 1.34 ]
Pohlandt 1992 11/63 13/74 18.7 % 0.99 [ 0.48, 2.06 ]
Subtotal (95% CI) 286 299 91.8 % 0.92 [ 0.66, 1.28 ]Total events: 53 (Long IT), 60 (Short IT)
Heterogeneity: Chi2 = 0.09, df = 2 (P = 0.96); I2 =0.0%
Test for overall effect: Z = 0.49 (P = 0.63)
2 No muscle relaxation
Spahr 1980 4/36 5/33 8.2 % 0.73 [ 0.21, 2.50 ]
Subtotal (95% CI) 36 33 8.2 % 0.73 [ 0.21, 2.50 ]Total events: 4 (Long IT), 5 (Short IT)
Heterogeneity: not applicable
Test for overall effect: Z = 0.50 (P = 0.62)
Total (95% CI) 322 332 100.0 % 0.91 [ 0.66, 1.24 ]Total events: 57 (Long IT), 65 (Short IT)
Heterogeneity: Chi2 = 0.22, df = 3 (P = 0.97); I2 =0.0%
Test for overall effect: Z = 0.61 (P = 0.54)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
28Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A98
Analysis 4.4. Comparison 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation),
Outcome 4 IVH (all grades).
Review: Long versus short inspiratory times in neonates receiving mechanical ventilation
Comparison: 4 Long vs Short IT (subgroup analysis by trials allowing use of muscle relaxation)
Outcome: 4 IVH (all grades)
Study or subgroup Long IT Short IT Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
1 Muscle relax allowed
Heicher 1981 13/51 12/51 46.9 % 1.08 [ 0.55, 2.14 ]
Subtotal (95% CI) 51 51 46.9 % 1.08 [ 0.55, 2.14 ]Total events: 13 (Long IT), 12 (Short IT)
Heterogeneity: not applicable
Test for overall effect: Z = 0.23 (P = 0.82)
2 No muscle relaxation
Spahr 1980 16/36 13/33 53.1 % 1.13 [ 0.65, 1.97 ]
Subtotal (95% CI) 36 33 53.1 % 1.13 [ 0.65, 1.97 ]Total events: 16 (Long IT), 13 (Short IT)
Heterogeneity: not applicable
Test for overall effect: Z = 0.42 (P = 0.67)
Total (95% CI) 87 84 100.0 % 1.11 [ 0.72, 1.71 ]Total events: 29 (Long IT), 25 (Short IT)
Heterogeneity: Chi2 = 0.01, df = 1 (P = 0.93); I2 =0.0%
Test for overall effect: Z = 0.46 (P = 0.65)
0.5 0.7 1 1.5 2
Favours Long IT Favours Short IT
W H A T ’ S N E W
Last assessed as up-to-date: 22 June 2003.
15 October 2008 Amended Converted to new review format.
H I S T O R Y
Protocol first published: Issue 4, 2003
Review first published: Issue 4, 2004
29Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A99
C O N T R I B U T I O N S O F A U T H O R S
Dr Kamlin and Dr Davis performed the literature search, assessed the methodological quality of eligible trials and extracted dataindependently. Dr Kamlin wrote the review and Dr Davis reviewed the manuscript.
D E C L A R A T I O N S O F I N T E R E S T
None
S O U R C E S O F S U P P O R T
Internal sources
• Royal Women’s Hospital, Melbourne, Australia.• Murdoch Children’s Research Institute, Melbourne, Australia.
External sources
• National Health and Medical Research Council, Australia.
I N D E X T E R M S
Medical Subject Headings (MeSH)
∗Inhalation; Hyaline Membrane Disease [∗complications]; Infant, Newborn; Infant, Premature; Intermittent Positive-Pressure Venti-lation [∗methods]; Randomized Controlled Trials as Topic; Respiratory Insufficiency [etiology; ∗therapy]; Time Factors
MeSH check words
Humans
30Long versus short inspiratory times in neonates receiving mechanical ventilation (Review)
Copyright © 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A100
ORIGINAL ARTICLE
Predicting successful extubation of very lowbirthweight infantsC O F Kamlin, P G Davis, C J Morley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
See end of article forauthors’ affiliations. . . . . . . . . . . . . . . . . . . . . . .
Correspondence to:Dr Kamlin, Division ofNewborn Services, RoyalWomen’s Hospital, 132Grattan Street, Carlton,Victoria 3053, Australia;[email protected]
Accepted16 December 2005Published Online First12 January 2006. . . . . . . . . . . . . . . . . . . . . . .
Arch Dis Child Fetal Neonatal Ed 2006;91:F180–F183. doi: 10.1136/adc.2005.081083
Objective: To determine the accuracy of three tests used to predict successful extubation of preterm infants.Study design: Mechanically ventilated infants with birth weight ,1250 g and considered ready forextubation were changed to endotracheal continuous positive airway pressure (ET CPAP) for threeminutes. Tidal volumes, minute ventilation (VE), heart rate, and oxygen saturation were recorded beforeand during ET CPAP. Three tests of extubation success were evaluated: (a) expired VE during ET CPAP; (b)ratio of VE during ET CPAP to VE during mechanical ventilation (VE ratio); (c) the spontaneous breathingtest (SBT)—the infant passed this test if there was no hypoxia or bradycardia during ET CPAP. The clinicalteam were blinded to the results, and all infants were extubated. Extubation failure was defined asreintubation within 72 hours of extubation.Results: Fifty infants were studied and extubated. Eleven (22%) were reintubated. The SBT was the mostaccurate of the three tests, with a sensitivity of 97% and specificity of 73% and a positive and negativepredictive value for extubation success of 93% and 89% respectively.Conclusion: The SBT used just before extubation of infants ,1250 g may reduce the number of extubationfailures. Further studies are required to establish whether the SBT can be used as the primary determinantof an infant’s readiness for extubation.
The adverse effects of prolonged endotracheal intubationon the preterm infant include bacterial colonisation,sepsis, subglottic injury, and bronchopulmonary dyspla-
sia.1 Thus extubation of ventilated infants as early as possibleis desirable.2 The decision to extubate is usually based onclinical assessment, blood gases, and ventilator settings.3
However, up to 40% of infants weighing ,1000 g who areextubated on these criteria require reintubation,4 suggestingthat the ability of clinicians to predict successful extubation islimited.5 Reintubation is destabilising, traumatic, and mayprolong the duration of mechanical ventilation and intensivecare.3 Accurate prediction of readiness for extubation ofpreterm infants may reduce morbidity.Pulmonary function tests to assess readiness for extubation
of adults5–7 and children8 9 have included measures of respira-tory muscle strength, respiratory drive, ventilatory reserve, lungfunction, and gas exchange. In newborn infants, pulmonaryfunction tests10–12 and invasive tests assessing musclestrength13 14 were less informative than gestational age.Respiratory drive is variable and sometimes inadequate in
preterm infants. Spontaneous expiratory minute ventilation(VE) of intubated infants provides an assessment ofrespiratory drive.15 Most modern ventilators measure VE
continuously and allow measurement during endotrachealcontinuous positive airway pressure (ET CPAP). Wilson andcolleagues15 calculated the VE ratio (ratio of ET CPAP VE toventilator VE) and found a ratio of .50% to have a positivepredictive value of 86% for extubation success. In arandomised controlled study of infants of all gestations, theVE ratio was shown to reduce the time to extubation but notimprove the rate of successful extubation.16
The aim of our study was to evaluate three tests ofreadiness for extubation in preterm infants. These wereconducted during three minutes of spontaneous breathing onET CPAP and included (a) measurement of VE, (b) calcula-tion of VE ratio, and (c) the spontaneous breathing test (SBT),which the infant was deemed to have passed if there was nohypoxia or bradycardia during ET CPAP.
METHODSSubjects/settingThe Royal Women’s Hospital, Melbourne, is a tertiaryperinatal centre with over 5000 deliveries per year. Thisstudy was conducted between June 2003 and October 2004.Infants with a birth weight of ,1250 g, ventilated using theDrager 8000 plus (software version 3; Drager Inc, Lubeck,Germany) for at least 24 hours, and thought by the clinicalteam to be ready for extubation were eligible. Infants wereweaned by either reducing delivered tidal volume to 3.5 ml/kg using assist-control or reducing ventilator rates to 20–30 breaths/min on synchronised intermittent mandatoryventilation. The inspiratory time was 0.3 second, the positiveend expiratory pressure 5–6 cm H2O, and inspired oxygenconcentration(40%. Extubation was considered if the infanthad satisfactory blood gases and was breathing above the setventilator rate. Sedation, when used, was stopped at least12 hours before extubation. Methylxanthines were usedbefore extubation or after extubation at the clinician’sdiscretion. Some infants were enrolled in a double blind trialof caffeine citrate versus placebo (the CAP trial).18 Theendotracheal tube was suctioned within two hours of thestudy.
ManoeuvreWhen the clinical team decided an infant was ready forextubation, the investigator switched the ventilator to ETCPAP at the same pressure as the positive end expiratorypressure setting. Airflow was measured using the sensor inthe ventilator circuit just proximal to the ET tube. Airflow,integrated expired tidal volume (Vte), and respiratory ratewere recorded at 200 Hz using Spectra software (GroveMedical, London, UK). Data were downloaded from the
Abbreviations: ET CPAP, endotracheal continuous positive airwaypressure; NIPPV, nasal intermittent positive pressure ventilation; ROC,receiver operating characteristic; SBT, spontaneous breathing test; VE,minute ventilation; Vte, expired tidal volume
F180
www.archdischild.com
A101
RS232 port of the ventilator during two minutes ofmechanical ventilation immediately before ET CPAP andthree minutes of ET CPAP. VE was the product of Vte andrespiratory rate. Expiratory VE was used because, when thereis an endotracheal tube leak, it is a more accurate measure ofthe volume of gas that enters the lungs than inspiratory VE.Mean respiratory rates, VE, and VE ratios were calculated. Afailed SBT was recorded if the infant had either a bradycardiafor more than 15 seconds and/or a fall in SpO2 below 85%despite a 15% increase in FIO2. At this point the study wasstopped, and ventilation restarted. Heart rate was displayedon HP M1205A Omnicare monitors (Hewlett-Packard,Andover, Massachusetts, USA) averaging every five beats.Clinical teams caring for the infants were not present
during the test and were blinded to the results. All infantswere extubated, regardless of the results of the test. Infantswere extubated to either nasal CPAP using Hudson prongs ornasal intermittent positive pressure ventilation (NIPPV) atthe clinician’s discretion.The primary study outcome was reintubation within
72 hours of extubation. The indications for reintubationwere specified a priori: (a) more than six episodes of apnoearequiring stimulation in six hours, or more than onesignificant episode of apnoea requiring bag and maskventilation; (b) respiratory acidosis (PaCO2 .65 mm Hg(.8.5 kPa) and pH,7.25); (c) FIO2- .0.60 to maintain SpO2in the target range (90–95%). The use of NIPPV was notconsidered a failure of extubation.The research and human research ethics committees of the
Royal Women’s Hospital, Melbourne approved the study, andinformed written consent was obtained before testing.
Statistical analysisReview of nursery data collected in 2002 revealed a 25%failure rate. We hypothesised that VE measured during thethree minute SBT would be lower in infants requiring
reintubation. A sample size of 50 was planned to detect adifference of one standard deviation in mean VE in the groupfailing extubation, assuming an overall mean (SD) VE of 300(50) ml/kg/min (power 80% and two tailed a of 0.05).Continuous outcomes were compared by Student’s t test
when normally distributed and by Mann-Whitney U testwhen skewed. Categorical data were assessed using theFisher two tailed exact test. p,0.05 was considered sig-nificant. The ability of respiratory variables to accuratelydiscriminate between successful and failed extubation wasassessed using receiver operating characteristic (ROC) curves.ROC curves are constructed by plotting the true positive rateon the y axis against the false positive rate on the x axis.Analysing the area under the curve determines the predictivevalue of a diagnostic test. A value (0.5 indicates nodiscriminatory value, and a value >0.8 suggests highpredictive value.19 Standard formulae were used to calculatesensitivity, specificity, and positive and negative predictivevalues and likelihood ratios. Statistical analyses wereperformed using SPSS version 11.5 (SPSS Inc, Chicago,Illinois, USA).
RESULTSDuring the study period, 251 babies with a birth weight,1250 g were admitted. Sixty five were ventilated from birthfor at least 24 hours. Parental consent was obtained for 50infants (31 male, 19 female). All had respiratory distresssyndrome. Forty one infants were weaned using synchro-nised intermittent mandatory ventilation and nine infants onassist-control. Extubation was successful in 39 (78%). Therewere no significant differences in gestational age, ventilatorsettings, internal diameter of ET tube, and birth or study
Table 1 Basic and clinical data on the study infants
Successfulextubation(n = 39)
Reintubationin 72 hours(n = 11)
Gestational age (weeks) 27.0 (1.7) 26.3 (2.2)Birth weight (g) 957 (215) 875 (232)Male/female 21/18* 10/1*Median age at study (days) 4 (1–45) 5 (1–21)Median weight at study (g) 1043 (490–1270) 864 (560–1145)ET tube internal diameter (2.5/3.0 mm) 27/12 10/1Weaning ventilator mode (SIMV/AC) 34/5 7/4FIO2 before ET CPAP (%) 24 (5) 26 (6)Mean airway pressure before ET CPAP (cm H2O) 7.1 (1.1) 7.3 (1.1)Ventilator rate before SBT (breaths/min) 29 (7) 32 (5)Respiratory support after extubation (NCPAP/NIPPV) 28/11 5/6
Data are mean (SD) or numbers unless otherwise stated. No significant differences (p,0.05) were found betweeninfants successfully extubated and those requiring reintubation. Male infants were more likely to be reintubated (*p= 0.03).ET CPAP, endotracheal continuous positive airway pressure; NCPAP, nasal continuous positive airway pressure;NIPPV, nasal intermittent positive pressure ventilation; SBT, spontaneous breathing test; SIMV, synchronisedintermittent mandatory ventilation; AC, assist-control.
2.0
1.5
1.0
0.5
0.0
VE ra
tio·
39Extubation
success
11Reintubation
Figure 1 Minute ventilation (VE)ratios (ratio of endotrachealcontinuous positive airwaypressure VE to ventilator VE) ininfants who were eithersuccessfully extubated orreintubated. Comparison ofmedians (Mann-Whitney) p =0.015.
Table 2 Variables predicting success or failure ofextubation by ROC curve analysis
VariableArea under theROC curve
95% confidenceintervals
Gestational age 0.61 0.39 to 0.89Birth weight 0.59 0.39 to0.79Postnatal age at study 0.55 0.36 to 0.74Mean airway pressure 0.56 0.37 to 0.74Tidal volume 0.57 0.38 to 0.76VE during ET CPAP 0.60 0.38 to 0.81VE ratios 0.74 0.56 to 0.92
ET CPAP, endotracheal continuous positive airway pressure; ROC,receiver operating characteristic.
Predicting extubation success in newborn infants F181
www.archdischild.com
A102
weights between successfully and unsuccessfully extubatedinfants (table 1). Most (64%; 7/11) reintubations occurred inthe first 24 hours (median 21 hours; range 1–71).Methylxanthines were used in 40 infants (23 before and 17after extubation). Two infants never received methyl-xanthines and remained extubated. Eight infants wereenrolled in the CAP study, and five were reintubated.Infants were extubated to either nasal CPAP (47) or NIPPV(three) as the primary mode of respiratory support. Anadditional 14 (28%) subsequently received NIPPV as rescuetreatment for apnoea. Significant apnoea accounted for eightof the 11 reintubations.The mean (SD) VE during ET CPAP was greater in
successfully extubated infants (314 (116) v 271 (113)), butthe difference did not reach statistical significance. Figure 1compares the VE ratios for the infants who were successfullyextubated with those reintubated. Although median VE ratioswere lower in infants who were reintubated (p = 0.015),there was considerable overlap between the groups.Table 2 presents the area under the ROC curves and the
95% confidence intervals. The inflection points of the ROCcurves were 220 ml/kg/min and 0.8 for VE and VE ratiosrespectively.Accuracy of the SBT was greater (higher sensitivity,
specificity, positive and negative predictive values) than theother two tests (table 3). Only three infants who passed theSBT were reintubated; one developed apnoea 71 hours afterextubation, one was intubated for respiratory acidosis and anincreasing oxygen requirement, and one developed stridorsoon after extubation.
DISCUSSIONOur study was designed to determine the accuracy of three tests(VE, VE ratios, and the SBT) in preterm infants judged ready forextubation on clinical grounds. In our study population, the SBTperformed better than both VE and VE ratios.This pragmatic observational study investigated infants
who were considered by clinicians to be ready for extubation.Our findings suggest that, if used to guide timing ofextubation, the SBT, although not infallible, may result infewer extubation failures and a greater proportion ofsuccessful extubations. False positive tests, such as thoseseen in two infants in this study (one with apnoea associatedwith subsequent sepsis and one with upper airway obstruc-tion), are unlikely to be anticipated by any test at the time ofextubation.Infants who are immature and recovering from respiratory
distress syndrome are at greatest risk of complications ofprolonged mechanical ventilation. However, it is not easy,even for experienced clinicians, to determine when theseinfants have sufficient respiratory drive to tolerate extuba-tion. It is not surprising that reported rates of reintubationvary from 15% to 40%.3 4 10 12–14 16 17 20 21 This variability islargely due to differences in study population. Only fourstudies investigated extubation of preterm infants.4 11 12 17 Welimited our study to infants ,1250 g because they represent
the majority of intubated infants and are at highest risk ofextubation failure. Accurate tests of readiness for extubationwould be a valuable addition to the clinical armamentarium.Studies examining respiratory muscle strength are invasiveand need expert application and interpretation. This andmeasurements of lung volumes after extubation limits theirclinical utility.13 14 21 Measurements of dynamic compliance,Vte, VE, and lung volumes have not provided clear thresholdvalues for reliable discrimination between extubation successand failure.3 6 8 10–12 14 20
The optimal duration of a trial of spontaneous breathingbefore extubation to determine an infant’s independentbreathing ability is uncertain. An adult study compared 30and 120 minutes and found no differences in safety or ratesof extubation success.22 Vento et al17 measured mean VE ofinfants ,1000 g during a two hour period on ET CPAP andconcluded that the test ‘‘could be useful’’ in identifying aninfant’s readiness for extubation. However, ET CPAP adds tothe resistance of the respiratory system23 leading to increasedwork of breathing, potentially jeopardising successful extu-bation. There is evidence that extubation after several hoursof ET CPAP is less successful than extubation from low rateventilation.24 Published studies evaluating VE ratios used10 minutes,15 16 whereas Veness-Meehan and coworkers6
measured Vte during one minute of ET CPAP. Beforeconducting this study, we observed that most infants whofailed a trial of ET CPAP did so in the first 90 seconds. Ourresults suggest that three minutes is long enough to identifymost infants who would fail extubation without riskingfatigue. The advantages and potential benefits of the SBT arethat it is easy to apply using standard monitoring devices inthe neonatal intensive care unit and simple to interpret.
Table 3 Summary statistics for the three variables used to predict extubation success
VariablePPV(%)
NPV(%)
Sens(%)
Spec(%) LR+ LR2
SBT 93 89 97 73 3.6 0.04VE (ET CPAP) .220 ml/kg/min
87 50 85 45 1.9 0.27
VE ratio >0.8 87 55 87 54 1.9 0.24
PPV, positive predictive value; NPV, negative predictive value; Sens, sensitivity; Spec, specificity; LR+, positivelikelihood ratio; LR2, negative likelihood ratio; SBT, spontaneous breathing test; VE, minute ventilation; ET CPAP,endotracheal continuous positive airway pressure.
What is already known on this topic
N Prolonged endotracheal intubation of preterm infants isassociated with short and long term morbidity
N At present there are no accurate tests for predictingreadiness for extubation, and up to 40% of extubatedinfants require reintubation
What this study adds
N Compared with measures of minute ventilation, inform-ation from a three minute period of endotracheal CPAPin infants considered ready for extubation is more likelyto identify those likely to require reintubation
N The spontaneous breathing test is simple to apply andeasier to interpret than other predictors of extubationsuccess
F182 Kamlin, Davis, Morley
www.archdischild.com
A103
In summary, we have shown that a simple three minuteSBT has a high negative predictive value when used in a unitwith access to both NCPAP and NIPPV to optimise manage-ment after extubation. Whether this test can be appliedearlier in the course of weaning to reduce the duration ofmechanical ventilation remains to be investigated.
ACKNOWLEDGEMENTSCOFK is supported by a Research Scholarship from the RoyalWomen’s Hospital, Melbourne, and PGD is supported by anAustralian National Health and Medical Research CouncilPractitioner Fellowship.
Authors’ affiliations. . . . . . . . . . . . . . . . . . . . .
C O F Kamlin, P G Davis, C J Morley, Division of Newborn Services,Royal Women’s Hospital, Melbourne, AustraliaP G Davis, C J Morley, University of Melbourne, Melbourne, AustraliaC J Morley, Murdoch Children’s Research Institute, Parkville, Melbourne,Australia
Competing interests: none declared
REFERENCES1 Bancalari E, Sinclair J. Effective care of the newborn infant, 1st ed. New York:
Oxford University Press, 1992.2 Halliday HL. Towards earlier neonatal extubation. Lancet 2000;355:2091–2.3 Fox WW, Schwartz JG, Shaffer TH. Successful extubation of neonates: clinical
and physiological factors. Crit Care Med 1981;9:823–6.4 Stefanescu BM, Murphy WP, Hansell BJ, et al. A randomized, controlled trial
comparing two different continuous positive airway pressure systems for thesuccessful extubation of extremely low birth weight infants. Pediatrics2003;112:1031–8.
5 Epstein SK. Decision to extubate. Intensive Care Med 2002;28:535–46.6 Veness-Meehan KA, Richter S, Davis JM. Pulmonary function testing prior to
extubation in infants with respiratory distress syndrome. Pediatr Pulmonol1990;9:2–6.
7 Yang KL, TobinMJ. A prospective study of indexes predicting the outcome of trialsof weaning from mechanical ventilation. N Engl J Med 1991;324:1445–50.
8 Khan N, Brown A, Venkataraman ST. Predictors of extubation success andfailure in mechanically ventilated infants and children. Crit Care Med1996;24:1568–79.
9 Baumeister BL, el-Khatib M, Smith PG, Blumer JL. Evaluation of predictors ofweaning from mechanical ventilation in pediatric patients. Pediatr Pulmonol1997;24:344–52.
10 Balsan MJ, Jones JG, Watchko JF, et al. Measurements of pulmonarymechanics prior to the elective extubation of neonates. Pediatr Pulmonol1990;9:238–43.
11 Smith J, Pieper CH, Maree D, et al. Compliance of the respiratory systemas a predictor for successful extubation in very-low-birth-weight infantsrecovering from respiratory distress syndrome. S Afr Med J1999;89:1097–102.
12 Kavvadia V, Greenough A, Dimitriou G. Prediction of extubation failure inpreterm neonates. Eur J Pediatr 2000;159:227–31.
13 Belani KG, Gilmour IJ, McComb C, et al. Preextubation ventilatorymeasurements in newborns and infants. Anesth Analg 1980;59:467–72.
14 Dimitriou G, Greenough A, Endo A, et al. Prediction of extubation failure inpreterm infants. Arch Dis Child Fetal Neonatal Ed 2002;86:F32–5.
15 Wilson BJ Jr, Becker MA, Linton ME, et al. Spontaneous minute ventilationpredicts readiness for extubation in mechanically ventilated preterm infants.J Perinatol 1998;18:436–9.
16 Gillespie LM, White SD, Sinha SK, et al. Usefulness of the minute ventilationtest in predicting successful extubation in newborn infants: a randomizedcontrolled trial. J Perinatol 2003;23:205–7.
17 Vento G, Tortorolo L, Zecca E, et al. Spontaneous minute ventilation is apredictor of extubation failure in extremely-low-birth-weight infants. J MaternFetal Neonatal Med 2004;15:147–54.
18 Schmidt B. Methylxanthine therapy for apnea of prematurity: evaluation oftreatment benefits and risks at age 5 years in the international Caffeine forApnea of Prematurity (CAP) trial. Biol Neonate 2005;88:208–13.
19 Beck JR, Shultz EK. The use of relative operating characteristic (ROC) curves intest performance evaluation. Arch Pathol Lab Med 1986;110:13–20.
20 Dimitriou G, Greenough A, Laubscher B. Lung volume measurementsimmediately after extubation by prediction of ‘‘extubation failure’’ inpremature infants. Pediatr Pulmonol 1996;21:250–4.
21 Dimitriou G, Greenough A. Computer assisted analysis of the chestradiograph lung area and prediction of failure of extubation from mechanicalventilation in preterm neonates. Br J Radiol 2000;73:156–9.
22 Esteban A, Alia I, Tobin MJ, et al. Effect of spontaneous breathing trialduration on outcome of attempts to discontinue mechanical ventilation.Spanish Lung Failure Collaborative Group. Am J Respir Crit Care Med1999;159:512–18.
23 LeSouef PN, England SJ, Bryan AC. Total resistance of the respiratory systemin preterm infants with and without an endotracheal tube. J Pediatr1984;104:108–11.
24 Davis PG, Henderson-Smart DJ. Extubation from low-rate intermittent positiveairways pressure versus extubation after a trial of endotracheal continuouspositive airways pressure in intubated preterm infants. Cochrane DatabaseSyst Rev 2001;(4):CD001078.
Predicting extubation success in newborn infants F183
www.archdischild.com
A104
A trial of spontaneous breathing to determine thereadiness for extubation in very low birth weightinfants: a prospective evaluationC O F Kamlin,1 P G Davis,1,2 B Argus,1 B Mills,1 C J Morley1,2,3
1 Division of Newborn Services,Royal Women’s Hospital,Melbourne, Australia;2 University of Melbourne,Australia; 3 Murdoch Children’sResearch Institute, Parkville,Melbourne, Australia
Correspondence to:Dr C O F Kamlin, Division ofNewborn Services, RoyalWomen’s Hospital, 132 GrattanStreet, Carlton, Victoria 3053,Australia; [email protected]
Accepted 24 November 2007Published Online First11 January 2008
ABSTRACTExtubation failure in premature infants is common. Aspontaneous breathing trial (SBT) was prospectivelyevaluated to determine timing of extubation. Comparedwith historical controls, infants were extubated atsignificantly higher ventilator rates and airway pressuresusing the SBT. No differences in rates of bronchopul-monary dysplasia or duration of ventilation were seen.
Endotracheal (ET) intubation and mechanical venti-lation remains the standard of care for managingrespiratory failure in very low birth weight (VLBW)infants. The adverse effects of prolonged endotra-cheal intubation include bacterial colonisation,sepsis, subglottic injury and bronchopulmonarydysplasia (BPD).1 The decision to extubate is usuallybased on clinical assessment. This is subjective andup to 40% of extubated infants require reinsertion oftheir ET tube.2 Reintubation is destabilising, trau-matic and may prolong the duration of mechanicalventilation.3 Objective criteria to identify the earliestopportunity to successfully extubate VLBW infantsare lacking.We have previously described and evaluated
respiratory drive in VLBW infants using a 3-minutespontaneous breathing trial (SBT) during ETcontinuous positive airway pressure (CPAP) beforeextubation.4 The test when applied once a clinicaldecision was made to extubate was highly sensitiveand appeared to identify infants for whomextubation was most likely to fail (a high negativepredictive value). The SBT was then incorporatedinto clinical practice at our hospital. The aim of ourcurrent study was to audit the effect of this changein policy on timing and success of extubation ofVLBW infants.
METHODSThe study was conducted between 1 June 2005 and30 June 2006 at The Royal Women’s Hospital,Melbourne and approved by the human researchand ethics committees as an audit of practice. Allventilated infants with a birth weight ,1250 gwere eligible for inclusion. Infants with lethal orchromosomal malformations or who died beforeextubation were excluded.Using the Drager Babylog 8000plus (Lubeck,
Germany), an SBT was performed when ventilatedinfants fulfilled entry criteria (set ventilator rate(45 bpm, volume guarantee (VG) (4.5 ml/kg orpeak inspiratory pressures (25 cm H2O, inspiredfractional oxygen (40%). This was repeated daily,typically on ward rounds until either extubation
criteria were met (‘‘successful SBT’’), unplannedextubation occurred or when the attending clin-ician made a decision to extubate (whichever wasthe sooner). Infants were ventilated and weanedusing either assist control (AC) or synchronisedintermittent mandatory ventilation (SIMV) ¡VGmodes. The ET CPAP pressure used was equivalentto the positive end expiratory pressure used inconventional ventilation (5–6 cm H2O). A failedSBT was recorded if the infant had bradycardia(,100/min) for .15 seconds and/or the SpO2 fellbelow 85% despite a 15% increase in fractionalinspired oxygen (FiO2) during the period of ETCPAP. At this point the SBT was discontinued andventilation restarted. Infants who passed the SBTwere extubated unless the attending cliniciansdisagreed. Non-compliance of the clinician withthe results of the SBT was recorded.The primary study outcome was reintubation
within 72 hours of extubation. The indications forreintubation were specified a priori: (a) more thansix episodes of apnoea requiring stimulation in6 hours, or more than one significant episode ofapnoea requiring bag and mask ventilation; (b)respiratory acidosis (PaCO2 .65 mm Hg (.8.5 kPa)and pH ,7.25); (c) FiO2 .0.60 to maintain SpO2 inthe target range (90–94%). BPD was defined assupplemental O2 or CPAP requirement at36 weeks’ corrected gestational age. Data werecompared with a cohort of VLBW infants admittedto the same institution in 2002, before the SBT waspart of clinical practice.
Power calculation and statistical analysisDuring a 2-year period (2003–4), the rate ofreintubation was 24%. Given the skewed nature ofthe duration of ventilation, the mean (SD) logtransformed duration of ventilation for VLBWinfants was 2.0 (0.7), corresponding to a geometricmean of 100 hours (102). We hypothesised a reduc-tion of 50% in the duration of assisted ventilationcompared with historical controls using the SBT todrive extubation (geometric mean of 50 hours, logtransformed =1.7). Eighty-seven patients in eachgroup were required to achieve 80% power and an aerror of 0.05. Statistical analyses were performedusing SPSS, version 11.5 (Cary, NJ, USA).
RESULTSOne hundred and eighty infants were studied (90SBT and 90 controls). Table 1 demonstrates that thebaseline variables were well matched between thetwo groups. However, fewer infants in the SBTcohort were intubated in the delivery room. During
Short report
Arch Dis Child Fetal Neonatal Ed 2008;93:F305–F306. doi:10.1136/adc.2007.129890 F305
A105
the latter period (SBT), VG ventilation was more commonly usedcompared with the control population (94% vs 26%) and therewas a shift of weaning ventilation modes at the time ofextubation from SIMV (93% in controls) to AC (81% in SBTcohort).The mean (SD) number of SBTs applied to infants was 3 (2).
Clinicians’ compliance with the SBT was excellent (97%). Six(7%) infants self-extubated after a failed SBT, three of whomwerereintubated. The sensitivity of the SBT to predict successfulextubation was 83%. The overall duration of mechanicalventilation was not significantly different (table 2). However,infants were extubated at significantly higher ventilator rates andmean airway pressure (Paw) when the SBT was applied. Durationof CPAP was significantly higher in the SBT cohort, althoughthere was no difference in the duration of all forms of ventilatorysupport or the incidence of BPD.
DISCUSSIONWe showed previously that the SBT applied to infants judgedby their clinician as being ready for extubation was accurate inpredicting success or failure. This audit of a change in clinicalpractice was performed in order to evaluate the safety andeffectiveness of the SBT to drive extubation of VLBW infants.Although using historical controls limits the number ofconclusions that can be drawn from the study, there weretwo potentially harmful outcomes that we wished to exclude.First, the period of endotracheal intubation could be prolongedif clinicians waited for infants to pass the SBT when they mightotherwise have decided to extubate earlier. Alternatively, byencouraging extubation sooner than the clinician might havedecided, more infants might have required reintubation.We have shown that invasivemechanical ventilatory support in
VLBW infants can be safely discontinued using the SBT atsignificantly higher Paw and set ventilator rates with similarreintubations compared with clinical decision-making. A clinicaltrial, randomising infants to have extubation guided by SBTversus clinical decision-making alone is therefore feasible. Fromour data, the median duration of ventilation for VLBW infants is,3 days and a very large sample size would be needed todemonstrate a significant reduction in duration of mechanicalventilation.Limitations of our study design are highlighted by the
differences seen in the two tables. While the patient demographicsand indicators of illness severity are similar, the data demonstrate
changes in the ventilator management of VLBW infants withinone institution. Significantly fewer infants are being intubated inthe delivery room and ventilator weaning modes have changed,with greater use of ACwith VG. Extubating infants from a higherPaw in the SBT group may be due changes in ventilator modebetween the two time periods. The proportion of supportedbreaths during AC would be greater than infants received onSIMV. Our data suggest it is safe to wean and extubate from ACmode with a mean Paw of 7.2 cm H2O.Our findings are consistent with reports in paediatric patients,
where standardised protocols for weaning compared with clinicaljudgment confer no benefit on duration of ventilation and rates ofextubation success.5 A single randomised study in newborninfants comparing defined weaning criteria based on minutevolumes with clinical assessment found a shorter duration inweaning from randomisation but in an underpowered study,more extubations failed with the intervention.6
In conclusion a 3-minute SBT to determine readiness toextubate VLBW infants is safe and as effective as clinicaldecision-making alone. Whether this mode of weaning can reduceduration of all respiratory support, BPD and length of stay needsfurther evaluation in a prospective randomised controlled study.
Funding: COFK is supported by a research scholarship from the Royal Women’s Hospital,Melbourne and PGD is supported by an Australian National Health and Medical ResearchCouncil practitioner fellowship.
Competing interests: None.
Ethics approval: Approved by the human research and ethics committees of The RoyalWomen’s Hospital, Melbourne, Australia
REFERENCES1. Bancalari E, Sinclair J. Mechanical ventilation. In: Sinclair J, Bracken M, eds. Effective
care of the newborn infant. New York: Oxford University Press, 1992:200–20.2. Stefanescu BM, Murphy WP, Hansell BJ, et al. A randomized, controlled trial
comparing two different continuous positive airway pressure systems for thesuccessful extubation of extremely low birth weight infants. Pediatrics2003;112:1031–8.
3. Veness-Meehan KA, Richter S, Davis JM. Pulmonary function testing prior toextubation in infants with respiratory distress syndrome. Pediatr Pulmonol 1990;9:2–6.
4. Kamlin CO, Davis PG, Morley CJ. Predicting successful extubation of very lowbirthweight infants. Arch Dis Child Fetal Neonat Ed 2006;91:F180–3.
5. Randolph AG, Wypij D, Venkataraman ST, et al. Effect of mechanical ventilatorweaning protocols on respiratory outcomes in infants and children: a randomizedcontrolled trial. JAMA 2002;28:2561–8.
6. Gillespie LM, White SD, Sinha SK, et al. Usefulness of the minute ventilation test inpredicting successful extubation in newborn infants: a randomized controlled trial.J Perinatol 2003;23:205–7.
Table 1 Demographic data, comorbidities of the spontaneous breathingtrial (SBT) cohort and of the historical controls
Data SBT (n=90) Control (n= 90) p Value
Birth weight* (g) 861 (229) 914 (189) 0.09
Gestational age* (weeks) 27 (2) 27 (2) 0.43
Multiple births (%) 34 (38) 38 (42) 0.12
Antenatal steroids (%) 72 (80) 70 (78) 0.72
Gender (male/female) 39/51 43/47 0.55
Ventilated from birth (%) 54 (60) 77 (86) ,0.001
Time from birth to intubation(hours)
0 (0–658) 0 (0–261) 0.001
Surfactant (%) 81 (90) 84 (93) 0.42
PDA treated (%) 37 (41) 33 (37) 0.54
Air leak (%) 7 (8) 10 (11) 0.44
HFOV (%) 16 (18) 13 (14) 0.54
Postnatal steroids (%) 7 (8) 6 (7) 0.77
Data are either mean (SD)* and analysed with independent t tests or number /proportions analysed by x2 analysis.HFOV, high-frequency oscillatory ventilation; PDA, patent ductus arteriosus.
Table 2 Pre-extubation ventilation parameters and outcomes followingextubation in both cohorts (spontaneous breathing trial (SBT) andhistorical controls)
ParametersSBT(n=90)
Control(n= 90) p Value
Duration of first period of ETventilation (days)
2.6 (0.2–48.8) 2.7 (0.1–47.0) 0.86
Ventilator rate at extubation(breaths/min)*
42 (6) 27 (8) ,0.001
MAP at extubation (cm H2O)* 7.2 (1.4) 6.5 (0.7) ,0.001
FiO2 at extubation* 0.23 (0.05) 0.23 (0.03) 0.31
Number of extubationsuccesses (%)
70 (78) 65 (72) 0.49
Days of nasal CPAP 24.4 (0.9–103.8) 15.3 (0.0–75.4) 0.01
Total duration of all forms ofassisted ventilation (days)
31.2 (1.1–141.0) 23.8 (0.3–125.6) 0.16
Number of cases of BPD (%) 32 (36) 32 (36) 1.00
Data are either proportions, mean (SD)* and analysed by x2 analysis/independent t-tests or median (range) and analysed using Mann–Whitney tests.BPD, bronchopulmonary dysplasia; CPAP, continuous positive airway pressure; ET,endotracheal; FiO2, fractional inspired oxygen; MAP, mean arterial pressure.
Short report
F306 Arch Dis Child Fetal Neonatal Ed 2008;93:F305–F306. doi:10.1136/adc.2007.129890
A106
Minerva Access is the Institutional Repository of The University of Melbourne
Author/s:Kamlin, Camille Omar Farouk
Title:Respiratory support of newborn infants in the delivery room and neonatal intensive care
Date:2010
Citation:Kamlin, C. O. F. (2010). Respiratory support of newborn infants in the delivery room andneonatal intensive care. Doctorate, Medicine, Dentistry & Health Sciences - Obstetrics andGynaecology (RWH), The University of Melbourne.
Persistent Link:http://hdl.handle.net/11343/35421
Terms and Conditions:Terms and Conditions: Copyright in works deposited in Minerva Access is retained by thecopyright owner. The work may not be altered without permission from the copyright owner.Readers may only download, print and save electronic copies of whole works for their ownpersonal non-commercial use. Any use that exceeds these limits requires permission fromthe copyright owner. Attribution is essential when quoting or paraphrasing from these works.
